NewYour coding agent can read the release notes before it upgrades.Set up the MCP server →
PyPI
PyPI, #4898 most downloaded on PyPI
A unified framework for machine learning with time series
Latest stable
1.3.0
Last release 13 days ago
08 Nov 2026
Release timing varies
gaps range from 1 weeks to 7 months
Nearly every release is documented
notes for 60 of the last 60 stable releases
1 version withdrawn
withdrawn after publishing
8 years old
105 releases · first in 2019
Drag across the ruler to compare, or tap a release to open it.
Users should also note completed deprecations and consult notes below.
Feature release with Python 3.15 compatibility.
Users should also note completed deprecations and consult notes below.
Please see our changelog for a description of all changes.
pandas 3.1.X compatibility (#11649) @fkiraly, @yash-sangwancategory tag for estimator categorization, e.g., "machine_learning", "foundation_model" (#11417) @jgyasutransformers 5 (#11237) @yftx293sktime framework.sktime framework, as Python 3.10 reached end-of-life.pandas 3.1.X is now supported by the sktime framework. The pandas upper bound has been raised to <3.2.polars (data container soft dependency) upper bound has been updated to <3.0.numba (computation soft dependency) upper bound has been updated to <0.69category tag for categorization, e.g., "machine_learning", "foundation_model". This allows to search for models in a specific category via all_estimators.shorthands tag, which can specify short names or aliases for specific instances of the object. These shorthands can be accessed via the craft utility.Benchmark analyzers are now first-class citizens, and inspectable via the registry, or craftable via craft. The type shorthand is benchmark_analyzer.
Time series detectors (BaseDetector`), classifiers (BaseClassifier), and regressors (``BaseRegressor) now have an extensible pretrain mechanism for handling pretrained models.
The pretraining API for detectors, classifiers, and regressors is not fully mature; model capabilities will be expanded in future releases.
From version 1.3.0, forecasters no longer universally provide "refitting " update capability by default and as a consequence no longer store all data seen so far in self._y and/or self._X.
Forecasters with an available "update" mode can now be queried via the capability:update tag.
To replicate the former default behaviour of "memorize all data and refit" in an upwards compatible way, users can use the UpdateRefitsEvery compositor from sktime.forecasting.stream.
In order to test future post-1.3.0 behaviour, users can set the config remember_data to False, via the .set_config(remember_data=False) (prior to forecaster use).
In tuners TSCGridSearchCV and TSRGridSearchCV, deprecated parameters n_jobs and pre_dispatch have been removed.
Configuration parameters for the parallelization backend should be passed via backend_params in the future.
@7524yash7524, @aakash1552005, @AbdulAliMamnun, @amarjaleelbanbhan, @amotl, @aryamanDutta, @chetana1410, @Dhruv-Kumar-Singh, @dhruvgupta2109, @direkkakkar319-ops, @Faakhir30, @fkiraly, @gthay, @Harshdeep-creator, @jgyasu, @KrishChadha001, @kritikhatri, @leandrepi, @markksantos, @MayankSharma-2812, @Me-Priyank, @nahcol10, @nikolas-sapa, @Nischal1425, @PiyushKumar74110, @PragnyaKhandelwal, @PredictiveManish, @RobKuebler, @S-h-u-b-h-1, @saptadeep12, @shivamkumxr, @shrohan2003, @shubham5080, @srupat, @SumanGouda, @Utkarshhs, @Viranshul-P, @Vivek1106-04, @VividhDesign, @yarnabrina, @yash-sangwan, @yftx293
Full Changelog: https://github.com/sktime/sktime/compare/v1.2.0...v1.3.0
Your coding agent can read these notes before it upgrades. Set up the MCP server →
Python 3.15 compatibility, pandas 3.1.X compatibility (11649) fkiraly, yash-sangwan
further reduction of estimator memory usage and stored model sizes (10802) Faakhir30
category tag for estimator categorization, e.g., "machine_learning", "foundation_model" (11417) jgyasu
TimeH2O forecasting foundation model (11694) KrishChadha001
Tabby forecasting foundation model (11691) saptadeep12
TimerS1 forecasting foundation model compatibility with transformers 5 (11237) yftx293
New forecasting metrics: MeanAbsoluteErrorPercentage (MAE%) and mean forecast bias (11637, 11582) Nischal1425, Utkarshhs
Python 3.15 is now supported by the sktime framework.
Python 3.10 is no longer supported by the sktime framework, as Python 3.10 reached end-of-life.
pandas 3.1.X is now supported by the sktime framework. The pandas upper bound has been raised to <3.2.
polars (data container soft dependency) upper bound has been updated to <3.0.
numba (computation soft dependency) upper bound has been updated to <0.69
All objects now can have a category tag for categorization, e.g., "machine_learning", "foundation_model". This allows to search for models in a specific category via all_estimators.
All objects now can have a shorthands tag, which can specify short names or aliases for specific instances of the object. These shorthands can be accessed via the craft utility.
Benchmark analyzers are now first-class citizens, and inspectable via the registry, or craftable via craft. The type shorthand is benchmark_analyzer.
Time series detectors (BaseDetector), classifiers (BaseClassifier), and regressors (BaseRegressor) now have an extensible pretrain mechanism for handling pretrained models.
The pretraining API for detectors, classifiers, and regressors is not fully mature; model capabilities will be expanded in future releases.
From version 1.3.0, forecasters no longer universally provide "refitting " update capability by default and as a consequence no longer store all data seen so far in self._y and/or self._X. Forecasters with an available "update" mode can now be queried via the capability:update tag. To replicate the former default behaviour of "memorize all data and refit" in an upwards compatible way, users can use the UpdateRefitsEvery compositor from sktime.forecasting.stream. In order to test future post-1.3.0 behaviour, users can set the config remember_data to False, via the .set_config(remember_data=False) (prior to forecaster use).
In tuners TSCGridSearchCV and TSRGridSearchCV, deprecated parameters n_jobs and pre_dispatch have been removed.
Configuration parameters for the parallelization backend should be passed via backend_params in the future.
[ENH] all_estimators to allow list-of-str for filter_tags; improve docstring (8252) nahcol10
[ENH] Environment Managers for Isolating Runtime Environments (11021) jgyasu
[ENH] dedicated namespace hook in registry (11316) fkiraly
[ENH] category tag for object categorization (11417) jgyasu
[ENH] namespace plugin for craft utility (11232) fkiraly
[ENH] Refactor _PretrainedCloner and _StateAtMixin to a common base for reuse with other estimator types, along with a _PretrainMixin for common pretrain operations (11534) srupat
[ENH] Merge _PretrainMixin and _StateAtMixin and delete _StateAtMixin (11629) srupat
[ENH] shorthands in craft (11627) fkiraly
[ENH] Add benchmark analyzers to the registry (11720) jgyasu
[ENH] MeanAbsoluteErrorPercentage (MAE%) forecasting metric (11637) Nischal1425
[ENH] Mean Forecast Bias metric (11582) Utkarshhs
[ENH] Add benchmark analyzers to the registry (11720) jgyasu
[ENH] Fix pandas 3.1.0 compatibility issues in forecasting and tests (11649) yash-sangwan
[ENH] Make VectorizedDF stateless (10802) Faakhir30
[ENH] test that check and convert do not mutate inputs (11609) Harshdeep-creator
[ENH] class-based convert_dict registry, migrate Hierarchical and Proba conversions to BaseConverter (11581)Ayushagrawal-cse, PragnyaKhandelwal
[ENH] golden output tests for TafsutForecaster (11351) AbdulAliMamnun
[ENH] category tag for object categorization (11417) jgyasu
[ENH] Add golden output regression test and fix router dtype for FalconTSTForecaster (11337) shivamkumxr
[ENH] minor fixes to tags (11529) fkiraly
[ENH] Support transformers 5 generation in TimerS1 (11237) yftx293
[ENH] Fix pandas 3.1.0 compatibility issues in forecasting and tests (11649) yash-sangwan
[ENH] TimeH2O forecasting foundation model (11694) KrishChadha001
[ENH] Tabby forecasting foundation model (11691) saptadeep12
[ENH] golden output tests for MIRAForecaster (11751) PragnyaKhandelwal
[ENH] Fix false-positive StatsForecastMSTL forecasting-horizon test (11479) PiyushKumar74110
[ENH] replace SARIMAX in an unrelated test by YfromX to reduce soft dependencies (11758) fkiraly
[ENH] Refactor: remove duplicate PyTorch base class (11706) srupat
[ENH] BaseDetector pretrain API (11206) yash-sangwan
[ENH] DummyPretrainedAnomalies with strategy=rate|pattern|timefreq text (11302) yash-sangwan
[ENH] Add test parameter coverage for MLPNetwork (11578) S-h-u-b-h-1
[ENH] allow arbitrary str from torch.optim, torch.nn and torch.optim.lr_scheduler in the pytorch base classes (11151) srupat
[ENH] Unify BaseDeepClassifierPyTorch and BaseDeepRegressorTorch to a common mixin adapter (11171) srupat
[ENH] Migrate gru & grufcn classifier from old _pytorch.py to new _base_torch.py interface (11426) srupat
[ENH] Migrate ConvTimeNet estimators to new PyTorch base interface (11325) srupat
[ENH] Migrate mvts_transformer classifier from old _pytorch.py to new _base_torch.py interface (11360) srupat
[ENH] in classifiers and regressors, restrict pyts version to last working (11674) fkiraly
[ENH] Pretrain capability for BaseClassifier and BaseRegressor (11596) srupat
[ENH] Add golden output regression test and fix router dtype for FalconTSTForecaster (11337) shivamkumxr
[ENH] allow arbitrary str from torch.optim, torch.nn and torch.optim.lr_scheduler in the pytorch base classes (11151) srupat
[ENH] fix _build_network signature in deep learning torch base classes (10955) srupat
[ENH] in DummyRegressor replace nested_univ with df-list / numpy3D (11436) shubham5080
[ENH] Unify BaseDeepClassifierPyTorch and BaseDeepRegressorTorch to a common mixin adapter (11171) srupat
[ENH] Migrate ConvTimeNet estimators to new PyTorch base interface (11325) srupat
[ENH] Pretrain capability for BaseClassifier and BaseRegressor (11596) srupat
[ENH] Add second parameter configuration for SupervisedIntervals (11381) amarjaleelbanbhan
[ENH] temporarily skip failing tests from 11495 (11496) fkiraly
[ENH] xfail test_catch22_wrapper_on_basic_motions due to change in feature calculation (11669) fkiraly
[ENH] Speed up KinematicFeatures row reductions (11707) dhruvgupta2109
[ENH] test that __all__ of all sktime modules lists names of module attributes (11418) Me-Priyank
[ENH] remove stale test skips for eight estimators whose skipped tests pass (11562) shrohan2003
[ENH] soft dependency free second test parameter sets, remove test_get_test_params_coverage skips for 14 estimators (11624) shrohan2003
[DOC] Fix links on "Community Showcase" page (11268) amotl
[DOC] make roadmap link also list closed roadmap issues (11279) fkiraly
[DOC] Add TimeSeriesAgglomerativeClustering usage example (11380) amarjaleelbanbhan
[DOC] minimal fix to docstring of run_test_module_changes (11394) fkiraly
[DOC] Fix typo in SFAFast docstring: "feature selectiona" → "feature selection" (9585) MayankSharma-2812
[DOC] Use integer fh shorthand in user guide and serialization docs (11445) aakash1552005
[DOC] minor improvements to load_basic_motions docstring (11465) fkiraly
[DOC] example in docstring of test transformer _TransformChangeNInstances (11498) fkiraly
[DOC] improve description of some values in category tag API reference (11530) fkiraly
[DOC] miscellaneous documentation improvements and changes (11446) fkiraly
[DOC] minor improvements for metrics API docs (11642) fkiraly
[DOC] Link developer installation card to contributor setup (11664) gthay
[DOC] CutoffSplitter - fix test window formula, add fold diagram (11464) 7524yash7524
[DOC] Add usage example for ShapeletLearningClassifierPyts (11678) dhruvgupta2109
[DOC] fix dead links in docs, notebooks and docstrings (11726) markksantos
[DOC] use integer fh in transformations module docstrings (11747) kritikhatri
[DOC] use integer fh in utils, param_est, base and pipeline docstrings (11760) kritikhatri
[MNT] [Dependabot](deps-dev): Update sphinx-gallery requirement from <0.22.0 to <0.23.0 (11266) dependabot[bot]
[MNT] all-contributors update (11303) github-actions[bot]
[MNT] Update core dependency upper bounds - scikit-base<1.3.0 and numba<0.69 (11391) aryamanDutta
[MNT] raise skpro bound to skpro<2.17 (11396) fkiraly
[MNT] add PragnyaKhandelwal to all-contributors (10036) PragnyaKhandelwal
[MNT] Bump github/codeql-action from 4.38.1 to 4.38.3 (11466) dependabot[bot]
[MNT] 1.3.0 deprecations - remove legacy benchmarking framework (11545) fkiraly
[MNT] Bump github/codeql-action from 4.38.3 to 4.39.0 (11567) dependabot[bot]
[MNT] all-contributors update (11620) github-actions[bot]
[MNT] updated the polars upper bound to <3.0 to unblock CI (11602) Utkarshhs
[MNT] Bump github/codeql-action from 4.39.0 to 4.39.1 (11641) dependabot[bot]
[MNT] 1.3.0 deprecations - flip remember_data config to False in all objects (11634) fkiraly
[MNT] 1.3.0 deprecations - remove deprecated fixture aliases and test_fixture_deprecation_module (11461) Utkarshhs
[MNT] Remove legacy 1.3.0 coverage name-swapping hack in Pipeline (11586) Utkarshhs
[MNT] raise pandas bound to pandas<3.2.0 (11633) fkiraly
[MNT] enable python 3.15 and newest scikit-base, skpro versions (11373) fkiraly
[MNT] all-contributors update (11670) github-actions[bot]
[MNT] migrate project.license to SPDX expression (PEP 639) (11705) kritikhatri
[MNT] add kritikhatri to all-contributors (11724) kritikhatri
[MNT] all-contributors update (11741) github-actions[bot]
[MNT] all-contributors update (11433) github-actions[bot]
[MNT] all-contributors update (11515) github-actions[bot]
[MNT] Add KrishChadha001 to all-contributors (11744) KrishChadha001
[MNT] Bump github/codeql-action from 4.39.1 to 4.39.2 (11721) dependabot[bot]
[MNT] Register Shubhaang Kataruka as sktime code contributor (11612) S-h-u-b-h-1
[MNT] remove outdated _contrib module (11757) fkiraly
[MNT] bump python lower bound to 3.11 (11450) yarnabrina
[MNT] 1.3.0 deprecations - Remove deprecated parallel parameters and bridge functions in GridSearch (11431) Utkarshhs
[MNT] 1.3.0 deprecations and change actions (11762) fkiraly
[BUG] fix median metrics failing with list or tuple horizon_weight (11370) VividhDesign
[BUG] Fix TypeError in sktime.catalogues.classification star import (11411) KrishChadha001
[BUG] fix recently merged MeanBias (11667) fkiraly
[BUG] fix default evaluate_by_index of forecasting metrics crashing with sample_weight (11702) PragnyaKhandelwal
[BUG] Fix MeanSquaredError.evaluate_by_index returning wrong values for RMSE (9575) PredictiveManish
[BUG] Fix TheilU2 confusing broadcast error when len(y_train) < sp (11731) Dhruv-Kumar-Singh
[BUG] fix SquaredDistrLoss using log_pdf instead of pdf (11750) PragnyaKhandelwal
[BUG] fix nested_univ target dtype with pandas 3 (11444) SumanGouda
[BUG] fix index levels in convert_Hierarchical_to_Panel (11523) PragnyaKhandelwal
[ENH] fix MCRecursiveProbaReductionForecaster use of self._cur_y and make its use in _get_expected_pred_idx explicit (11386) fkiraly
[BUG] fix predict_quantiles default from predict_interval mixing up variables for multivariate y (11424) VividhDesign
[BUG] fix compatibility of BaseProbaForecaster with new VectorizedDF (11524) fkiraly
[BUG] fit trend forecasters on time since the series start (11520) chetana1410
[BUG] fix T0Forecaster contract test failures (11539) KrishChadha001
[BUG] In ReconcilerForecaster, fix residual calculation before forecaster update (11644) SumanGouda
[BUG] fix numerical instability of PolynomialTrendForecaster for large time stamps (11472) shrohan2003
[BUG] replace eval with getattr in vendored Chronos create_tokenizer (11645) Me-Priyank
[BUG] fix NormalHedgeEnsemble crash when no estimator has positive regret (11428) Viranshul-P
[BUG] treat non-finite scores as failed blueprints in AutoResearchForecaster (11682) RobKuebler
[BUG] fix AutoResearchForecaster update not updating best_forecaster_ (11686) Vivek1106-04
[BUG] fix FlowStateForecaster fails estimator suite tests on MacOS (11532) KrishChadha001
[BUG] fix hypertrees forecaster with pandas3 (11560)RugvedBane, PiyushKumar74110
[BUG] fix NaN and Inf detection logic in global pooling reduction forecasters (#11521) (11701) aakash1552005
[BUG] Preserve the MPS availability wrapper in PatchTST test parameters (11723) aryamanDutta
[BUG] fix SyntaxError in BaseDetector.segments_to_change_points doctest (9925) direkkakkar319-ops
[BUG] fix _build_network signature in deep learning torch base classes (10955) srupat
[BUG] fix CNTCClassifier custom objects method name for save/load (10971) srupat
[BUG] BaggingClassifier: stop reseeding the process-global numpy RNG (11307) nikolas-sapa
[BUG] fix TEASER by replacing sklearn cross-validation with sktime-native fit_predict_proba call (11440) fkiraly
[BUG] fix early classifier _state handling and add some missing doctest test skips (11753) fkiraly
[BUG] remove spurious decision_function from classification and clustering scenarios (11442) fkiraly
[BUG] fix SeasonalDummiesOneHot for quarterly, weekly and hourly data (11183) yash-sangwan
[BUG] fix multirocket first-order features computation (11292) leandrepi
[BUG] Fix Imputer._init_ method in "forecaster" equality check (11690) Dhruv-Kumar-Singh
[BUG] Fix timestamp midpoints in TimeBinAggregate on pd.DateTimeIndex (11737) KrishChadha001
[BUG] remove spurious decision_function from classification and clustering scenarios (11442) fkiraly
7524yash7524, aakash1552005, AbdulAliMamnun, amarjaleelbanbhan, amotl, aryamanDutta, chetana1410, Dhruv-Kumar-Singh, dhruvgupta2109, direkkakkar319-ops, Faakhir30, fkiraly, gthay, Harshdeep-creator, jgyasu, KrishChadha001, kritikhatri, leandrepi, markksantos, MayankSharma-2812, Me-Priyank, nahcol10, nikolas-sapa, Nischal1425, PiyushKumar74110, PragnyaKhandelwal, PredictiveManish, RobKuebler, S-h-u-b-h-1, saptadeep12, shivamkumxr, shrohan2003, shubham5080, srupat, SumanGouda, Utkarshhs, Viranshul-P, Vivek1106-04, VividhDesign, yarnabrina, yash-sangwan, yftx293
Users should also note completed deprecations and consult notes below.
Feature release.
Users should also note completed deprecations and consult notes below.
Please see our changelog for a description of all changes.
sktime package size, smaller estimator memory footprint (#10459, #10892) @faakhir30, @fkiralypandas 3 support (#11056, #11153) @yash-sangwanMCRecursiveProbaReductionForecaster (#9242) @marrovfable::ARIMA forecaster (via rpy2) (#8641) @ericjbhypertrees-forecasting: HyperTreeARForecaster (#11076) @oberoir080evaluate utility now can run benchmarks for pretrain-able forecasters (#10560) @jgyasuSavitzkyGolayTransformer, HilbertTransformer, and WaveletPacketTransformer (#10810) @ved197338EventTPR, MeanDetectionOffset and FalseAlarmRate for live event detection (#11197) @yash-sangwancraft utility for deserializing estimator specifications now has a safe argument which prevents arbitrary serialized code in the estimator definition. Only a restricted syntax with estimator names in sktime is allowed.tbats package is no longer maintained and has been vendored into sktime.libs.
The sktime vendor is also updated for compatibility with newest numpy and scikit-learn.From version 1.3.0, forecasters will no longer universally provide "refitting " update capability by default and as a consequence no longer store all data seen so far in self._y and/or self._X.
Forecasters with an available "update" mode can now be queried via the capability:update tag.
To replicate the former default behaviour of "memorize all data and refit" in an upwards compatible way, users can use the UpdateRefitsEvery compositor from sktime.forecasting.stream.
In order to test future post-1.3.0 behaviour, users can set the config remember_data to False, via the .set_config(remember_data=False) (prior to forecaster use).
In line with the forecaster update default change, from version 1.3.0, self._y and self._X will no longer store all data seen so far by default.
Users who wish to retain the previous behavior in an upwards compatible way can set the config remember_data to True via the .set_config(remember_data=True) (prior to forecaster use).
Deprecation of the capability:global_forecasting tag is complete, it has been removed entirely.
Users should use the capability:pretrain tag instead.
@8rulerstar, @AanchalGupta1162, @AbdulAliMamnun, @AbiramiR-27, @adan-shahid, @adity1raut, @AH64-dll, @AlejandroCoronadoN, @AlvaroBalbin, @amarjaleelbanbhan, @Anai-Guo, @andoriyaprashant, @anupamkr1708, @archittmittal, @aryamanDutta, @Ashish-Kumar-Dash, @Ayushagrawal-cse, @baremetaldevx86, @Belagum, @blackorange-mega, @Boubker10, @chala2001, @Chetansahney, @chintam-dhanush, @chrisprown902-wq, @CloseChoice, @coding-cosmos, @dannymaaz, @DebD-max, @DebojitNath, @dhairya-motta, @dhruvb2028, @direkkakkar319-ops, @dongwonmoon, @ericjb, @Faakhir30, @fathirramadhan-web, @felipebridge, @fkiraly, @fnhirwa, @gabrielemidulla, @geetu040, @godarrenw, @goyaladitya05, @gthay, @Harshavardhan-28, @hasanfaesal, @hdimer, @HuzaifaAbdulRehman, @InnoxCodes, @JamesBoardman27, @jbbqqf, @jgyasu, @jvsch, @kajal-jotwani, @Keykyrios, @KingLizard1020, @LouisDeconinck, @LunarScoop, @luziyi123448-gif, @maarcosrmz, @marrov, @mccool1010, @Me-Priyank, @MGPOCKY, @Muhammad-Rebaal, @NAME-ASHWANIYADAV, @nicolasdmolina, @Nikunjsaini07, @Nischal1425, @oberoir080, @OfficialAbhinavSingh, @Partharsid, @PiyushKumar74110, @Pragati5-DEBUG, @pruthvipatill, @Ridadata, @RobKuebler, @sabasiddique1, @Saswatsusmoy, @Sharon-study, @shivamlalakiya, @shubham5080, @siddharth7113, @Siddhartha23i, @snoopuppy582, @Spicy-source, @srupat, @sudo-muneeb, @tarundb2005-create, @TayfurYldz, @Tharun-10Dragneel, @Udayan853, @Valentino-source-dev, @ved197338, @vedantag17, @WAHIB-EL-KHADIRI, @webzuweb, @wunianze666-netizen, @YadavAkash96, @yash-sangwan, @zelihaguven
Full Changelog: https://github.com/sktime/sktime/compare/v1.1.0...v1.2.0
better memory management: smaller sktime package size, smaller estimator memory footprint ( #10459 , #10892 ) @faakhir30 , @fkiraly
pandas 3 support ( #11056 , #11153 ) @yash-sangwan
New foundation model forecasters: Tafsut, TimesFM3, TiRex-2, TFC-T0 ( #10447 , #10786 , #10976 , #11053 ) @aryamanDutta , @hasanfaesal , @siddharth7113 , @yash-sangwan
Recursive reduction using Monte-Carlo to probabilistic tabular regressors, MCRecursiveProbaReductionForecaster ( #9242 ) @marrov
CRAN fable::ARIMA forecaster (via rpy2 ) ( #8641 ) @ericjb
interface to hypertrees-forecasting : HyperTreeARForecaster ( #11076 ) @oberoir080
forecasting evaluate utility now can run benchmarks for pretrain -able forecasters ( #10560 ) @jgyasu
new transformations: SavitzkyGolayTransformer , HilbertTransformer , and WaveletPacketTransformer ( #10810 ) @ved197338
EventTPR , MeanDetectionOffset and FalseAlarmRate for live event detection ( #11197 ) @yash-sangwan
Hierarchical/agglomerative clustering for time series ( #9187 ) @Muhammad-Rebaal
pandas 3.X is now supported by the sktime framework.
numpy upper bounds have been updated to <2.6 .
scikit-learn upper bounds have been updated to <1.10 .
joblib upper bounds have been updated to <1.7 .
The craft utility for deserializing estimator specifications now has a safe argument which prevents arbitrary serialized code in the estimator definition. Only a restricted syntax with estimator names in sktime is allowed.
From version 1.3.0, forecasters will no longer universally provide “refitting “ update capability by default and as a consequence no longer store all data seen so far in self._y and/or self._X . Forecasters with an available “update” mode can now be queried via the capability:update tag. To replicate the former default behaviour of “memorize all data and refit” in an upwards compatible way, users can use the UpdateRefitsEvery compositor from sktime.forecasting.stream .
In line with the forecaster update default change, from version 1.3.0, self._y and self._X will no longer store all data seen so far by default. Users who wish to retain the previous behavior in an upwards compatible way can set the config remember_data to True via the .set_config(remember_data=True) (prior to forecaster use).
The tbats package is no longer maintained and has been vendored into sktime.libs . The sktime vendor is also updated for compatibility with newest numpy and scikit-learn .
From version 1.3.0, forecasters will no longer universally provide “refitting “ update capability by default and as a consequence no longer store all data seen so far in self._y and/or self._X . Forecasters with an available “update” mode can now be queried via the capability:update tag. To replicate the former default behaviour of “memorize all data and refit” in an upwards compatible way, users can use the UpdateRefitsEvery compositor from sktime.forecasting.stream . In order to test future post-1.3.0 behaviour, users can set the config remember_data to False , via the .set_config(remember_data=False) (prior to forecaster use).
Deprecation of the capability:global_forecasting tag is complete, it has been removed entirely. Users should use the capability:pretrain tag instead.
[ENH] migrate sktime_version tag to _BaseTag class ( #10805 , #10812 ) @Spicy-source , @Partharsid
[ENH] ensure sktime_version tag is set dynamically to ensure inclusion in object serialization ( #10831 ) @fkiraly
[ENH] migrate skip-inverse-transform tag to _BaseTag class ( #10811 ) @DebojitNath
[ENH] remove task and duplicate distribution_type from ESTIMATOR_TAG_REGISTER ( #10859 ) @AanchalGupta1162
[ENH] migrate scitype:X tag to _BaseTag class ( #10876 ) @AanchalGupta1162
[ENH] migrate symmetric tag to BaseTag class ( #10847 ) @aryamanDutta
[ENH] migrate scitype:instancewise tag to _BaseTag class ( #10886 ) @anupamkr1708
[ENH] migrate capability:multithreading tag to _BaseTag class ( #10848 ) @wunianze666-netizen
[ENH] migrate split_type tag to _BaseTag class ( #10937 ) @Valentino-source-dev
[ENH] migrate X-y-must-have-same-index from tuple to class based registry ( #10981 ) @PiyushKumar74110
[ENH] “safe” mode in craft utility ( #11086 ) @fkiraly
[ENH] migrate scitype:y tag from tuple to _BaseTag class ( #10985 ) @gabrielemidulla
[ENH] make string coercion in scikit-learn adapter pandas 3 compatible ( #11137 ) @fkiraly
[ENH] remove leftover X-y-must-have-same-index tuple from ESTIMATOR_TAG_REGISTER ( #11149 ) @shubham5080
[ENH] migrate test_get_test_params_coverage skips from _config to class tags ( #11193 ) @Keykyrios
[ENH] migrate test skip lists to estimator tags (Task 2) ( #10663 ) @sudo-muneeb
[ENH] centralize lookup module skip configs in registry , and fix libs not skipped in sklearn estimator lookup ( #11222 ) @fkiraly
[ENH] migrate test_doctest_examples skips from _config to class tags ( #11221 ) @Me-Priyank
[ENH] use ME for monthly date_range in tests on pandas 2.2+ ( #11122 ) @yash-sangwan
[ENH] migrate remember_data tag to _BaseTag class ( #11115 ) @KingLizard1020
[ENH] update scikit-learn _version_bridge to also deal with force_all_finite / ensure_all_finite rename ( #11181 ) @fkiraly
[ENH] migrate learning_type tag to _BaseTag class ( #10806 ) @Nikunjsaini07
[ENH] migrate split_hierarchical tag to _BaseTag class ( #10935 ) @Valentino-source-dev
[ENH] migrate capability:missing_values:removes tag to class ( #10909 ) @luziyi123448-gif
[ENH] migrate split_series_uses tag to _BaseTag class ( #10938 ) @Valentino-source-dev
[ENH] raise on incomplete benchmark spec instead of returning empty results ( #10764 ) @yash-sangwan
[ENH] Adapt forecasting evaluate to the new pretrain API (global benchmarking) ( #10560 ) @jgyasu
[ENH] Make BenchmarkingResults public ( #10873 ) @jgyasu
[ENH] use modern frequency aliases in dataset loaders ( #11070 ) @yash-sangwan
[ENH] replace MTYPE_SOFT_DEPS and SCITYPE_REGISTER by programmatic lookup ( #9812 ) @goyaladitya05
[ENH] golden output tests for MomentFMForecaster ( #10685 ) @geetu040
[ENH] golden output tests for TotoForecaster ( #10684 ) @geetu040
[ENH] golden output tests for Toto2Forecaster ( #10683 ) @geetu040
[ENH] golden output tests for MOIRAIForecaster ( #10686 ) @geetu040
[ENH] LagLlamaForecaster - add test registry reference to vendor library ( #10719 ) @fkiraly
[ENH] golden output tests for HFTransformersForecaster ( #10751 ) @geetu040
[ENH] MOIRAI Forecaster caching fixes ( #10460 ) @vedantag17
[ENH] Toto forecaster: Pretrain addition and lazy load fixes ( #10462 ) @vedantag17
[ENH] interface T0 forecaster from The Forecasting Company ( #10447 ) @siddharth7113
[ENH] golden output tests for MantisForecaster ( #10736 ) @geetu040
[ENH] Add padding_mask param to TinyTimeMixerForecaster to select padding-mask semantics ( #10735 ) @geetu040
[ENH] Add second test parameter sets for RandomIntervalClassifier , DummyRegressor , and StackingForecaster ( #10785 ) @NAME-ASHWANIYADAV
[ENH] Add second test parameter sets for reduction forecasters ( #10679 ) @NAME-ASHWANIYADAV
[ENH] add golden-output regression test for TimesFM2Forecaster ( #10674 ) @Nischal1425
[ENH] Add dedicated _update method to Croston forecaster ( #10819 ) @DebD-max
[ENH] CRAN fable::ARIMA forecaster (via rpy2 ) ( #8641 ) @ericjb
[ENH] Move DynamicFactor test skip to estimator tag ( #10861 ) @LunarScoop
[ENH] Adapt forecasting evaluate to the new pretrain API (global benchmarking) ( #10560 ) @jgyasu
[ENH] Add test parameter sets to estimators ( #9527 ) @maarcosrmz
[ENH] ForecastingHorizon pandas 3 compatibility ( #10868 ) @yash-sangwan
[ENH] golden output tests for LagLlamaForecaster ( #10682 ) @geetu040
[ENH] minor fixes: replace deprecated internal import path, remove stray newlines in docstring ( #10960 ) @fkiraly
[ENH] better isolation of CINNForecaster imports ( #11001 ) @fkiraly
[ENH] better isolation of MomentFMForecaster imports ( #11000 ) @fkiraly
[ENH] better isolation of PykanForecaster imports ( #11039 ) @fkiraly
[ENH] better isolation of HFTransformersForecaster imports ( #11038 ) @fkiraly
[ENH] golden output tests for WindFMForecaster ( #11043 ) @AbdulAliMamnun
[ENH] TiRex2Forecaster interfacing the TiRex-2 foundation model ( #10786 ) @yash-sangwan
[ENH] update deprecated frequency aliases in test_fh for pandas 3 ( #11056 ) @yash-sangwan
[ENH] Recursive reduction using Monte-Carlo to probabilistic tabular regressors, MCRecursiveProbaReductionForecaster ( #9242 ) @marrov
[ENH] move HypertreeNetARForecaster to module and adds API reference ( #11103 ) @fkiraly
[ENH] interface to hypertrees-forecasting : HyperTreeARForecaster ( #11076 ) @oberoir080
[ENH] use modern frequency aliases in tests ( #11121 ) @yash-sangwan
[ENH] add scikit-learn bound to tbats forecasters, remove tbats from tutorials ( #11113 ) @fkiraly
[ENH] Vendor tbats into sktime.libs and fix incompatibilities with newer numpy , scipy , scikit-learn ( #11150 ) @DebojitNath
[ENH] pandas 3 compatibility patches to estimators ( #11153 ) @yash-sangwan
[ENH] replace no longer compatible scikit-learn make_pipeline with sktime make_pipeline in reducer tests ( #11155 ) @fkiraly
[ENH] fix compatibility failures in test_reduce ( #11189 ) @Ayushagrawal-cse
[ENH] remove cross-module imports in classification , regression , transformation , and clustering modules ( #11213 ) @fkiraly
[ENH] Removal of self._X and self._y from BaseForecaster ( #10459 ) @Faakhir30
[ENH] remove tests:skip_all from SARIMAX and FreshPRINCE ( #11233 ) @8rulerstar
[ENH] better isolation of soft dependencies in TimerS1 vendor module ( #11224 ) @fkiraly
[ENH] TimesFM 3 multivariate foundation model forecaster ( #10976 ) @hasanfaesal
[ENH] forecaster capability:update tag and dedicated API tests ( #11247 ) @fkiraly
[ENH] pretrain support for TimeMoE foundation model ( #11729 ) @faakhir30
[ENH] add missing license statement to libs.tbats vendor ( #11157 ) @fkiraly
[ENH] fix tbats incompatibility with latest scikit-learn ( #11165 ) @fathirramadhan-web
[ENH] add TBATS vendor tests ( #11170 ) @PiyushKumar74110
[ENH] golden output tests for ChronosForecaster ( #10692 ) @geetu040
[ENH] golden output tests for Moirai2Forecaster ( #10687 ) @geetu040
[ENH] golden output tests for Chronos2Forecaster ( #10677 ) @geetu040
[ENH] golden output tests for KronosForecaster ( #10997 ) @dongwonmoon
[ENH] golden output tests for FlowStateForecaster ( #11048 ) @AbdulAliMamnun
[ENH] golden output tests for PatchTSTForecaster ( #10941 ) @Nischal1425
[ENH] consistent automatic CPU/GPU placement for foundation-model forecasters ( #11167 ) @geetu040
[ENH] Tafsut foundation model forecaster ( #11053 ) @aryamanDutta
[ENH] in AutoTS , make silent verbose=-2 parameter the default ( #11263 ) @fkiraly
[ENH] add NeuralProphet soft dependency bound for pandas<3 ( #11259 ) @amarjaleelbanbhan
[ENH] clean up soft dependency isolation patterns in networks module ( #11088 ) @fkiraly
[ENH] Revert #11088 soft dependency isolation as it compromises pickleability, use _safe_import instead ( #11231 ) @fkiraly
[ENH] Split param_est.stationarity._arch into separate estimator submodules ( #10552 ) @NAME-ASHWANIYADAV
[ENH] Fix Singular matrix in ImpulseResponseFunction VECM tests ( #10977 ) @DebojitNath
[ENH] Add second test param sets for ElasticEnsemble , DistFromAligner ( #10815 ) @NAME-ASHWANIYADAV
[ENH] replace SubLOF in tests with dummy detectors ( #11152 ) @fkiraly
[ENH] make SubLOF pandas 3 compatible ( #11136 , #11141 ) @fkiraly , @TayfurYldz
[ENH] EventTPR , MeanDetectionOffset and FalseAlarmRate for live detection ( #11197 ) @yash-sangwan
[ENH] remove detection.all based aliasing in detection root imports ( #11261 ) @fkiraly
[ENH] FreshPRINCE classifier - unskip tests, fix tags ( #10711 ) @fkiraly
[ENH] Add second test param sets for ElasticEnsemble , DistFromAligner ( #10815 ) @NAME-ASHWANIYADAV
[ENH] Add second test parameter sets for RandomIntervalClassifier, DummyRegressor, and StackingForecaster ( #10785 ) @NAME-ASHWANIYADAV
[ENH] Add _get_keras_custom_objects method to base classifier classes for loading deep learning models involving custom layers ( #10637 ) @srupat
[ENH] Torch-based FCN time series classifier and regressor ( #9489 ) @kajal-jotwani
[ENH] Fix mutable default arguments in ConvTimeNet backbone ( #10835 ) @Boubker10
[ENH] add testing tags to deep learning base classes so changes in them trigger tests in child classes ( #10944 ) @fkiraly
[ENH] torch -based CNTC time-series classifier and regressor ( #10920 ) @srupat
[ENH] torch based cntc time series classifier and regressor ( #9397 ) @fnhirwa
[ENH] Migrate WEASEL test skip config from tests._config to estimator tags ( #10904 ) @JamesBoardman27
[ENH] Move MACNNClassifier test skip config to tags ( #9814 ) @archittmittal
[ENH] Add doctest examples to deep learning classes ( #10956 ) @srupat
[ENH] Add second test parameter set for SummaryClassifier , SupervisedTimeSeriesForest ( #10871 ) @Keykyrios
[ENH] clean up tags for deep learning classifiers and regressors ( #10953 ) @fkiraly
[ENH] Add second test parameter sets for ContractableBOSS , MUSE , and WEASEL ( #10934 ) @NAME-ASHWANIYADAV
[ENH] skip ConvTimeNetClassifier test failures on main until resolved ( #11032 ) @fkiraly
[ENH] Add second test parameter sets for TSFreshClassifier ( #10999 ) @oberoir080
[ENH] native grid search for TSCGridSearchCV and TSRGridSearchCV ( #10818 ) @yash-sangwan
[ENH] Make LSTMFCN TensorFlow conv layers dynamic via filter_sizes and kernel_sizes ( #11078 ) @nicolasdmolina
[ENH] pandas 3 compatibility patches to estimators ( #11153 ) @yash-sangwan
[ENH] add upper scikit-learn bound to ElasticEnsemble ( #11154 ) @fkiraly
[ENH] add scikit-learn compatibility bound to TimeSeriesForestClassifier , remove as test case ( #11176 ) @fkiraly
[ENH] register TimeSeriesForestClassifier tests in tag ( #11179 ) @fkiraly
[ENH] replace TimeSeriesForestClassifier in more second party tests ( #11182 ) @fkiraly
[ENH] remove scikit-learn inheritance from TimeSeriesForestClassifier and TimeSeriesForestRegressor ( #11178 ) @InnoxCodes
[ENH] further scikit-learn 1.8.X compatibility fixes ( #11185 ) @fkiraly
[ENH] skip sktime in scikit-learn compatibility tests ( #11186 ) @fkiraly
[ENH] add scikit-learn<1.8 bounds to pyts estimators ( #11210 ) @fkiraly
[ENH] remove cross-module imports in classification , regression , transformation , and clustering modules ( #11213 ) @fkiraly
[ENH] remove tests:skip_all from SARIMAX and FreshPRINCE ( #11233 ) @8rulerstar
[ENH] ResNet network migration from tf to torch ( #10762 ) @srupat
[ENH] Add _get_keras_custom_objects method to base classifier classes for loading deep learning models involving custom layers ( #10637 ) @srupat
[ENH] Hierarchical/Agglomerative Clustering for Time Series ( #9187 ) @Muhammad-Rebaal
[ENH] Add second test parameter set for TimeSeriesKMedoids ( #10870 ) @Harshavardhan-28
[ENH] remove cross-module imports in classification , regression , transformation , and clustering modules ( #11213 ) @fkiraly
[ENH] Add second test parameter sets for RandomIntervalClassifier, DummyRegressor, and StackingForecaster ( #10785 ) @NAME-ASHWANIYADAV
[BUG] Add _get_keras_custom_objects method to base classifier classes for loading deep learning models involving custom layers ( #10637 ) @srupat
[ENH] restore test skips in CntcRegressor ( #10875 ) @fkiraly
[ENH] Torch-based FCN time series classifier and regressor ( #9489 ) @kajal-jotwani
[ENH] add testing tags to deep learning base classes so changes in them trigger tests in child classes ( #10944 ) @fkiraly
[ENH] torch -based CNTC time-series classifier and regressor ( #10920 ) @srupat
[ENH] torch based cntc time series classifier and regressor ( #9397 ) @fnhirwa
[ENH] Add doctest examples to deep learning classes ( #10956 ) @srupat
[ENH] clean up tags for deep learning classifiers and regressors ( #10953 ) @fkiraly
[ENH] native grid search for TSCGridSearchCV and TSRGridSearchCV ( #10818 ) @yash-sangwan
[ENH] Make LSTMFCN TensorFlow conv layers dynamic via filter_sizes and kernel_sizes ( #11078 ) @nicolasdmolina
[ENH] remove scikit-learn inheritance from TimeSeriesForestClassifier and TimeSeriesForestRegressor ( #11178 ) @InnoxCodes
[ENH] remove cross-module imports in classification , regression , transformation , and clustering modules ( #11213 ) @fkiraly
[ENH] ResNet network migration from tf to torch ( #10762 ) @srupat
[ENH] ClaSPTransformer second test parameter set ( #8868 ) @Udayan853
[ENH] address numpy.trapz deprecation and fix test tags in related estimators ( #10727 ) @fkiraly
[ENH] Add second test parameter set for ClearSky ( #10788 ) @coding-cosmos
[ENH] Migrate skip-inverse-transform tag to _BaseTag class ( #10811 ) @DebojitNath
[ENH] Add second FeatureSelection test parameter set ( #10232 ) @snoopuppy582
[ENH] Add SavitzkyGolayTransformer , HilbertTransformer , and WaveletPacketTransformer ( #10810 ) @ved197338
[ENH] make FourierFeatures transformation compatible with pandas 3 frequency aliases ( #10895 ) @yash-sangwan
[ENH] make Imputer mean/median compatible with pandas 3 ( #10924 ) @yash-sangwan
[ENH] Add unconditional HolidayFeatures test parameters ( #11045 ) @Sharon-study
[ENH] make TimeSince tests pandas 3 compatible ( #11134 ) @fkiraly
[ENH] pandas 3 compatibility patches to estimators ( #11153 ) @yash-sangwan
[ENH] add scikit-learn<1.8 bounds to pyts estimators ( #11210 ) @fkiraly
[ENH] remove cross-module imports in classification , regression , transformation , and clustering modules ( #11213 ) @fkiraly
[ENH] Add second test parameter set for DilationMappingTransformer ( #10657 ) @OfficialAbhinavSingh
[ENH] Add second test parameter sets for PAAlegacy , SAXlegacy , and Hidalgo ( #10732 ) @NAME-ASHWANIYADAV
[ENH] Refactor test framework step 1: behaviour-preserving fixture rename (#10647) ( #10661 ):user: Faareh-Ahmed , @yash-sangwan
[ENH] Refactor test framework step 2: PackageConfig mixin and attribute alignment (#10647) ( #10807 ) @yash-sangwan
[ENH] Refactor test framework step 3: inherit fixture engine from skbase (#10647) ( #10862 ) @yash-sangwan
[ENH] Refactor test framework step 4: Inherit TestAllObjects and remove redundant test bodies ( #11079 ) @yash-sangwan
[ENH] add test that all classes have docstrings and at least one doctest example ( #10776 ) @fkiraly
[ENH] remove legacy EXCLUDE_ESTIMATORS mechanism from tests._config ( #11208 ) @fkiraly
[DOC] fix links for pytorch-forecasting ( #10689 ) @CloseChoice
[DOC] remove ARM troubleshooting section which is no longer applicable ( #10694 ) @fkiraly
[DOC] fix minor formatting issue in the mlflow integration docstrings ( #10777 ) @fkiraly
[DOC] add doctest example for DegreeDayFeatures ( #10817 ) @chintam-dhanush
[DOC] Add doctest example for TimeSeriesDBSCAN ( #10801 ) @Nikunjsaini07
[DOC] add usage example for Id transformer ( #10787 ) @baremetaldevx86
[DOC] Add doctest example for Catch22Wrapper #10795 ( #10799 ) @Nikunjsaini07
[DOC] add Python 3.14 to supported versions in get started guide ( #10830 ) @AH64-dll
[DOC] add doctest examples for five classes ( #10796 ) @baremetaldevx86
[DOC] fix DontUpdate docstring ( #10834 ) @fkiraly
[DOC] add doctest examples for SlidingWindowSegmenter ( #10836 ) @chintam-dhanush
[DOC] complete VAR docstring for method , missing , freq , dates ( #10698 ) @shivamlalakiya
[DOC] fix undocumented configs in various foundation models ( #10630 ) @Faakhir30
[DOC] Fix broken links and typos in ConditionalDeseasonalizer docs ( #10864 ) @CloseChoice
[DOC] Fix Outdated Dockerfile Example and Table in continuous_integration.rst ( #10877 ) @JamesBoardman27
[DOC] fix broken links ( #10865 ) @CloseChoice
[DOC] adds docstring example to PCATransformer ( #10869 ) @pruthvipatill
[DOC] add doctest example for DirectedChamfer (#10782) ( #10858 ) @Pragati5-DEBUG
[DOC] add doctest examples for DerivativeSlopeTransformer ( #10827 ) @chintam-dhanush
[DOC] Move network parameter docs to class docstrings (LSTMFCN, InceptionTime) ( #9521 ) @sabasiddique1
[DOC] Update links to contributor emoji key ( #10884 ) @JamesBoardman27
[DOC] Fix FABBA docstring parameter names to match the actual signature ( #10942 ) @godarrenw
[DOC] add doctest example for CutoffFhSplitter ( #10939 ) @Valentino-source-dev
[DOC] Add missing randomness tags to forecasting template ( #10889 ) @Siddhartha23i
[DOC] document property:randomness and capability:random_state in extension templates ( #10908 ) @chala2001
[DOC] document estimator serialization format ( #10596 ) @dannymaaz
[DOC] Update governance guidelines link in bug reporting documentation ( #10913 ) @JamesBoardman27
[DOC] add doctest example for FittedParamExtractor ( #10963 ) @OfficialAbhinavSingh
[DOC] fix capability:categorical_in_X tag description and incorrect tag defaults in tag registry ( #10854 ) @siddharth7113
[DOC] add doctest example for SqueezeHierarchy ( #11013 ) @chrisprown902-wq
[DOC] add doctest example for IgnoreX ( #11034 ) @oberoir080
[DOC] Fix broken links in vendored fracdiff README ( #11052 ) @blackorange-mega
[DOC] Add doctest example to SAXlegacy ( #11066 ) @adan-shahid
[DOC] add doctest example for DistFromAligner ( #11090 ) @LouisDeconinck
[DOC] add doctest example for MatrixProfileFeatures ( #11109 ) @shubham5080
[DOC] update README in sktime.libs ( #11160 ) @Ayushagrawal-cse
[DOC] add doctest example for SignatureKernel ( #11130 ) @shubham5080
[DOC] add doctest examples for dummy catalogues ( #11077 ) @HuzaifaAbdulRehman
[DOC] Add usage example to BoxCoxBiasAdjustedForecaster ( #9429 ) @YadavAkash96
[DOC] add doctest example for SupervisedIntervals ( #11205 ) @zelihaguven
[DOC] add docstring example to RBFForecaster ( #11235 ) @fkiraly
[DOC] add doctest examples for TimeBinAggregate ( #10826 ) @chintam-dhanush
[DOC] add doctest example for WindowSegmenter ( #10863 ) @chintam-dhanush
[DOC] add capability:multivariate and capability:pairwise to docs ( #10880 ) @AanchalGupta1162
[DOC] add docstring example for AutoTS forecaster ( #11262 ) @fkiraly
[DOC] add doctest examples for detection cost functions ( #10982 ) @jvsch
[MNT] Add CodeQL Analysis Workflow ( #8406 ) @fkiraly , @andoriyaprashant
[MNT] add name to release tag check step for readability ( #10708 ) @MGPOCKY
[MNT] all-contributors update ( #10700 ) @github-actions[bot]
[MNT] version upgrade for upper bounds of deps: numpy<2.6 , scikit-base<1.2.0 , numba<0.67 , and some soft deps ( #10723 ) @fkiraly
[MNT] deprecate the v1 benchmarking framework (#10464) ( #10636 ) @yash-sangwan
[MNT] all-contributors update ( #10820 ) @github-actions[bot]
[MNT] refactor test for compatibility with historical dependencies to matrix, add 2025 ( #10823 ) @fkiraly
[MNT] fix minor bugs in the test.yml CI file ( #10824 ) @fkiraly
[MNT] Update numba requirement from <0.67,>=0.53 to >=0.53,<0.68 ( #10860 ) @dependabot[bot]
[MNT] all-contributors update ( #10878 ) @github-actions[bot]
[MNT] remove extraneous material from package wheel ( #10892 ) @fkiraly
[MNT] Trigger website docs on stable releases ( #10638 ) @gthay
[MNT] all-contributors update ( #10946 ) @github-actions[bot]
[MNT] remove unnecessary from future imports for defunct python 3.9 and lower versions ( #11033 ) @fkiraly
[MNT] all-contributors update ( #11042 ) @github-actions[bot]
[MNT] Dummy init.py to avoid test fails in the CI ( #11075 ) @jgyasu
[MNT] Ensure skpro>=2 in CI ( #11081 ) @fkiraly
[MNT] fix skpro>=2 in notebook dependencies ( #11084 ) @PiyushKumar74110
[MNT] raise joblib bound to joblib<1.7 ( #11096 ) @fkiraly
[MNT] add historical 2026 snapshot for testing ( #11114 ) @fkiraly
[MNT] core dependency upper bound monitor ( #11095 ) @fkiraly
[MNT] Update skpro requirement from <2.15.0,>=2 to >=2,<2.16.0 ( #11132 ) @dependabot[bot]
[MNT] do not cancel test-historical-deps all runs if one run fails ( #11135 ) @fkiraly
[MNT] allow pandas 3 ( #9279 ) @fkiraly
[MNT] remove upper bounds from notebooks depset ( #11158 ) @fkiraly
[MNT] Update pytest-randomly requirement from <4.2,>=3.15 to >=3.15,<5.1 ( #11174 ) @dependabot[bot]
[MNT] raise scikit-learn bound to scikit-learn<1.10.0 ( #10726 ) @fkiraly
[MNT] update statsmodels upper bounds in pyproject.toml to statsmodels<0.16 ( #11112 ) @fkiraly
[MNT] all-contributors update ( #11094 ) @github-actions[bot]
[MNT] all-contributors update ( #11199 ) @github-actions[bot]
[MNT] Bump github/codeql-action from 4.37.3 to 4.38.1 ( #10747 , #10775 , #10849 , #10890 , #11133 , #11241 ) @dependabot[bot]
[MNT] Preserve MPS wrapper metadata for PyTorch 2.14 ( #11047 ) @PiyushKumar74110
[BUG] Fix unraised exceptions, minor logic and docstring issues ( #9373 ) @Ashish-Kumar-Dash
[BUG] ignore scikit-learn / sktime conftest modules in scikit-learn estimator registry lookup ( #11004 ) @geetu040
[BUG] in Pipeline , fix Step._fetch_input_data leaking transformer_names across edge groups ( pandas 3 ) ( #11142 ) @felipebridge
[BUG] make BaseObject.eq check the class, not just the params ( #10781 ) @yash-sangwan
[BUG] mean_squared_log_error : remove leftover debug print and stack trace ( #10816 ) @AlejandroCoronadoN
[BUG] Pin scipy version for neuralprophet ( #10738 ) @Faakhir30
[BUG] CiscoTSMForecaster predict_proba uses HistogramQPD ( #10554 ) @Saswatsusmoy
[BUG] Fix CiscoTSMForecaster._predict_proba inconsistency using HistogramQPD ( #10559 ) @AbiramiR-27
[BUG] Fix mutable default arguments in vendored LagLlama code ( #10733 ) @Ridadata
[BUG] Fix SCINetForecaster test failure in test_predict_series_name_preserved ( #10289 ) @dhairya-motta
[BUG] Fix unraised exceptions, minor logic and docstring issues ( #9373 ) @Ashish-Kumar-Dash
[BUG] set capability:categorical_in_X to False for forecasters whose backend requires numeric X ( #10855 ) @siddharth7113
[BUG] Fix NaiveVariance with forecasters raising NotImplementedError , such as ExponentialSmoothing on statsmodels 0.15+ ( #10958 ) @aryamanDutta
[BUG] exclude non-finite scores from AutoResearchForecaster blueprint selection ( #10965 ) @OfficialAbhinavSingh
[BUG] Fix random_state handling in statsmodels adapter ( #10972 ) @aryamanDutta
[BUG] fix VARMAX and DynamicFactor failing with integer index on statsmodels 0.15 ( #11119 ) @Belagum
[BUG] fix Chronos-Bolt long prediction length ( #11172 ) @Ayushagrawal-cse
[BUG] Greykite API conformance fixes ( #10755 ) @Faakhir30
[BUG] Fix SquaringResiduals breaking on multi-step fh ( #10740 ) @Faakhir30
[BUG] fix ARIMA and AutoARIMA out-of-sample prediction with statsmodels 0.15 for unsupported indices ( #11236 ) @mccool1010
[BUG] Mutable default argument in WindowWarp for LagLLama ( #10097 ) @direkkakkar319-ops
[BUG] Disable capability:non_contiguous_X for SARIMAX ( #10741 ) @Faakhir30
[BUG] Fix mutable defaults in LagLlama estimator ( #10879 ) @tarundb2005-create
[BUG] Preserve pretrained PatchTSMixer configuration ( #10921 ) @dhruvb2028
[BUG] fix off-by-one in PatchTST prediction_length derived from fh ( #10940 ) @Nischal1425
[BUG] fix incorrect pd.Series _predict return name in _PytorchForecastingAdapter ( #11253 ) @fkiraly
[BUG] fix B006 mutable default args in convtimenet backbones ( #10730 ) @WAHIB-EL-KHADIRI
[BUG] Fix mutable default arguments in ConvTimeNet backbone ( #10835 ) @Boubker10
[BUG] in DistFromAligner , correct symmetric tag ( #11110 ) @fkiraly
[BUG] ensure DistFromAligner treats symmetric property directly ( #11108 ) @webzuweb
[BUG] Resolve bugs in the TapNet Tensorflow implementation ( #9447 ) @srupat
[BUG] fix deepcopy of estimators using SeqSelfAttentionTorch ( #10929 ) @srupat
[BUG] fix optimizer passed as an instance not being bound to the network in BaseDeepClassifierPytorch ( #11035 ) @srupat
[BUG] remove moved ensemble classifiers from classification.compose.all ( #11104 ) @Anai-Guo
[BUG] fix ConvTimeNetClassifier deepcopy and pickling failures ( #11099 ) @adity1raut
[BUG] fix typo in DistFromAligner._transform causing TypeError when X2 is None ( #10841 ) @NAME-ASHWANIYADAV
[BUG] fix missing boolean negation in test_multioutput for TSR ( #10809 ):user: atikulmunna , @fkiraly
[BUG] Resolve bugs in the TapNet Tensorflow implementation ( #9447 ) @srupat
[BUG] fix deepcopy of estimators using SeqSelfAttentionTorch ( #10929 ) @srupat
[BUG] fix (batch, 1) output handling in BaseDeepRegressorTorch ( #10932 ) @srupat
[BUG] fix ResNetRegressor broken init ( #10949 ) @fkiraly
[BUG] Fix optimizer_kwargs silently ignored in BaseDeepRegressorTorch due to wrong guard condition ( #10926 ) @srupat
[BUG] reenable tests and fix tags for TSFreshFeatureExtractor ( #10715 ) @fkiraly
[BUG] fix off-by-one in TimeBinAggregate bin_mid index (#10173) ( #10214 ) @jbbqqf
[BUG] fix "bin_mid" index in TimeBinAggregate in integer case, add test coverage ( #10175 ):user: fkiraly , @Tharun-10Dragneel
[BUG] Fix: raise ValueError with correct message in RandomSamplesAugmenter (use f-string) ( #11026 ) @PiyushKumar74110
[BUG] fix PeakTimeFeature on panel and hierarchical input ( #10919 ) @Chetansahney
@8rulerstar , @AanchalGupta1162 , @AbdulAliMamnun , @AbiramiR-27 , @adan-shahid , @adity1raut , @AH64-dll , @AlejandroCoronadoN , @AlvaroBalbin , @amarjaleelbanbhan , @Anai-Guo , @andoriyaprashant , @anupamkr1708 , @archittmittal , @aryamanDutta , @Ashish-Kumar-Dash , @Ayushagrawal-cse , @baremetaldevx86 , @Belagum , @blackorange-mega , @Boubker10 , @chala2001 , @Chetansahney , @chintam-dhanush , @chrisprown902-wq , @CloseChoice , @coding-cosmos , @dannymaaz , @DebD-max , @DebojitNath , @dhairya-motta , @dhruvb2028 , @direkkakkar319-ops , @dongwonmoon , @ericjb , @Faakhir30 , @fathirramadhan-web , @felipebridge , @fkiraly , @fnhirwa , @gabrielemidulla , @geetu040 , @godarrenw , @goyaladitya05 , @gthay , @Harshavardhan-28 , @hasanfaesal , @hdimer , @HuzaifaAbdulRehman , @InnoxCodes , @JamesBoardman27 , @jbbqqf , @jgyasu , @jvsch , @kajal-jotwani , @Keykyrios , @KingLizard1020 , @LouisDeconinck , @LunarScoop , @luziyi123448-gif , @maarcosrmz , @marrov , @mccool1010 , @Me-Priyank , @MGPOCKY , @Muhammad-Rebaal , @NAME-ASHWANIYADAV , @nicolasdmolina , @Nikunjsaini07 , @Nischal1425 , @oberoir080 , @OfficialAbhinavSingh , @Partharsid , @PiyushKumar74110 , @Pragati5-DEBUG , @pruthvipatill , @Ridadata , @RobKuebler , @sabasiddique1 , @Saswatsusmoy , @Sharon-study , @shivamlalakiya , @shubham5080 , @siddharth7113 , @Siddhartha23i , @snoopuppy582 , @Spicy-source , @srupat , @sudo-muneeb , @tarundb2005-create , @TayfurYldz , @Tharun-10Dragneel , @Udayan853 , @Valentino-source-dev , @ved197338 , @vedantag17 , @WAHIB-EL-KHADIRI , @webzuweb , @wunianze666-netizen , @YadavAkash96 , @yash-sangwan , @zelihaguven
better memory management: smaller sktime package size, smaller estimator memory footprint (10459, 10892) faakhir30, fkiraly
pandas 3 support (11056, 11153) yash-sangwan
New foundation model forecasters: Tafsut, TimesFM3, TiRex-2, TFC-T0 (10447, 10786, 10976, 11053) aryamanDutta, hasanfaesal, siddharth7113, yash-sangwan
Recursive reduction using Monte-Carlo to probabilistic tabular regressors, MCRecursiveProbaReductionForecaster (9242) marrov
CRAN fable::ARIMA forecaster (via rpy2) (8641) ericjb
interface to hypertrees-forecasting: HyperTreeARForecaster (11076) oberoir080
forecasting evaluate utility now can run benchmarks for pretrain-able forecasters (10560) jgyasu
new transformations: SavitzkyGolayTransformer, HilbertTransformer, and WaveletPacketTransformer (10810) ved197338
EventTPR, MeanDetectionOffset and FalseAlarmRate for live event detection (11197) yash-sangwan
Hierarchical/agglomerative clustering for time series (9187) Muhammad-Rebaal
pandas 3.0.X is now supported by the sktime framework. The upper bound has been raised to <3.1.
numpy upper bounds have been updated to <2.6.
scikit-learn upper bounds have been updated to <1.10.
joblib upper bounds have been updated to <1.7.
The craft utility for deserializing estimator specifications now has a safe argument which prevents arbitrary serialized code in the estimator definition. Only a restricted syntax with estimator names in sktime is allowed.
From version 1.3.0, forecasters will no longer universally provide "refitting " update capability by default and as a consequence no longer store all data seen so far in self._y and/or self._X. Forecasters with an available "update" mode can now be queried via the capability:update tag. To replicate the former default behaviour of "memorize all data and refit" in an upwards compatible way, users can use the UpdateRefitsEvery compositor from sktime.forecasting.stream.
In line with the forecaster update default change, from version 1.3.0, self._y and self._X will no longer store all data seen so far by default. Users who wish to retain the previous behavior in an upwards compatible way can set the config remember_data to True via the .set_config(remember_data=True) (prior to forecaster use).
The tbats package is no longer maintained and has been vendored into sktime.libs. The sktime vendor is also updated for compatibility with newest numpy and scikit-learn.
From version 1.3.0, forecasters will no longer universally provide "refitting " update capability by default and as a consequence no longer store all data seen so far in self._y and/or self._X. Forecasters with an available "update" mode can now be queried via the capability:update tag. To replicate the former default behaviour of "memorize all data and refit" in an upwards compatible way, users can use the UpdateRefitsEvery compositor from sktime.forecasting.stream. In order to test future post-1.3.0 behaviour, users can set the config remember_data to False, via the .set_config(remember_data=False) (prior to forecaster use).
Deprecation of the capability:global_forecasting tag is complete, it has been removed entirely. Users should use the capability:pretrain tag instead.
[ENH] migrate sktime_version tag to _BaseTag class (10805, 10812) Spicy-source, Partharsid
[ENH] ensure sktime_version tag is set dynamically to ensure inclusion in object serialization (10831) fkiraly
[ENH] migrate skip-inverse-transform tag to _BaseTag class (10811) DebojitNath
[ENH] remove task and duplicate distribution_type from ESTIMATOR_TAG_REGISTER (10859) AanchalGupta1162
[ENH] migrate scitype:X tag to _BaseTag class (10876) AanchalGupta1162
[ENH] migrate symmetric tag to BaseTag class (10847) aryamanDutta
[ENH] migrate scitype:instancewise tag to _BaseTag class (10886) anupamkr1708
[ENH] migrate capability:multithreading tag to _BaseTag class (10848) wunianze666-netizen
[ENH] migrate split_type tag to _BaseTag class (10937) Valentino-source-dev
[ENH] migrate X-y-must-have-same-index from tuple to class based registry (10981) PiyushKumar74110
[ENH] "safe" mode in craft utility (11086) fkiraly
[ENH] migrate scitype:y tag from tuple to _BaseTag class (10985) gabrielemidulla
[ENH] make string coercion in scikit-learn adapter pandas 3 compatible (11137) fkiraly
[ENH] remove leftover X-y-must-have-same-index tuple from ESTIMATOR_TAG_REGISTER (11149) shubham5080
[ENH] migrate test_get_test_params_coverage skips from _config to class tags (11193) Keykyrios
[ENH] migrate test skip lists to estimator tags (Task 2) (10663) sudo-muneeb
[ENH] centralize lookup module skip configs in registry, and fix libs not skipped in sklearn estimator lookup (11222) fkiraly
[ENH] migrate test_doctest_examples skips from _config to class tags (11221) Me-Priyank
[ENH] use ME for monthly date_range in tests on pandas 2.2+ (11122) yash-sangwan
[ENH] migrate remember_data tag to _BaseTag class (11115) KingLizard1020
[ENH] update scikit-learn _version_bridge to also deal with force_all_finite / ensure_all_finite rename (11181) fkiraly
[ENH] migrate learning_type tag to _BaseTag class (10806) Nikunjsaini07
[ENH] migrate split_hierarchical tag to _BaseTag class (10935) Valentino-source-dev
[ENH] migrate capability:missing_values:removes tag to class (10909) luziyi123448-gif
[ENH] migrate split_series_uses tag to _BaseTag class (10938) Valentino-source-dev
[ENH] raise on incomplete benchmark spec instead of returning empty results (10764) yash-sangwan
[ENH] Adapt forecasting evaluate to the new pretrain API (global benchmarking) (10560) jgyasu
[ENH] Make BenchmarkingResults public (10873) jgyasu
[ENH] use modern frequency aliases in dataset loaders (11070) yash-sangwan
[ENH] replace MTYPE_SOFT_DEPS and SCITYPE_REGISTER by programmatic lookup (9812) goyaladitya05
[ENH] golden output tests for MomentFMForecaster (10685) geetu040
[ENH] golden output tests for TotoForecaster (10684) geetu040
[ENH] golden output tests for Toto2Forecaster (10683) geetu040
[ENH] golden output tests for MOIRAIForecaster (10686) geetu040
[ENH] LagLlamaForecaster - add test registry reference to vendor library (10719) fkiraly
[ENH] golden output tests for HFTransformersForecaster (10751) geetu040
[ENH] MOIRAI Forecaster caching fixes (10460) vedantag17
[ENH] Toto forecaster: Pretrain addition and lazy load fixes (10462) vedantag17
[ENH] interface T0 forecaster from The Forecasting Company (10447) siddharth7113
[ENH] golden output tests for MantisForecaster (10736) geetu040
[ENH] Add padding_mask param to TinyTimeMixerForecaster to select padding-mask semantics (10735) geetu040
[ENH] Add second test parameter sets for RandomIntervalClassifier, DummyRegressor, and StackingForecaster (10785) NAME-ASHWANIYADAV
[ENH] Add second test parameter sets for reduction forecasters (10679) NAME-ASHWANIYADAV
[ENH] add golden-output regression test for TimesFM2Forecaster (10674) Nischal1425
[ENH] Add dedicated _update method to Croston forecaster (10819) DebD-max
[ENH] CRAN fable::ARIMA forecaster (via rpy2) (8641) ericjb
[ENH] Move DynamicFactor test skip to estimator tag (10861) LunarScoop
[ENH] Adapt forecasting evaluate to the new pretrain API (global benchmarking) (10560) jgyasu
[ENH] Add test parameter sets to estimators (9527) maarcosrmz
[ENH] ForecastingHorizon pandas 3 compatibility (10868) yash-sangwan
[ENH] golden output tests for LagLlamaForecaster (10682) geetu040
[ENH] minor fixes: replace deprecated internal import path, remove stray newlines in docstring (10960) fkiraly
[ENH] better isolation of CINNForecaster imports (11001) fkiraly
[ENH] better isolation of MomentFMForecaster imports (11000) fkiraly
[ENH] better isolation of PykanForecaster imports (11039) fkiraly
[ENH] better isolation of HFTransformersForecaster imports (11038) fkiraly
[ENH] golden output tests for WindFMForecaster (11043) AbdulAliMamnun
[ENH] TiRex2Forecaster interfacing the TiRex-2 foundation model (10786) yash-sangwan
[ENH] update deprecated frequency aliases in test_fh for pandas 3 (11056) yash-sangwan
[ENH] Recursive reduction using Monte-Carlo to probabilistic tabular regressors, MCRecursiveProbaReductionForecaster (9242) marrov
[ENH] move HypertreeNetARForecaster to module and adds API reference (11103) fkiraly
[ENH] interface to hypertrees-forecasting: HyperTreeARForecaster (11076) oberoir080
[ENH] use modern frequency aliases in tests (11121) yash-sangwan
[ENH] add scikit-learn bound to tbats forecasters, remove tbats from tutorials (11113) fkiraly
[ENH] Vendor tbats into sktime.libs and fix incompatibilities with newer numpy, scipy, scikit-learn (11150) DebojitNath
[ENH] pandas 3 compatibility patches to estimators (11153) yash-sangwan
[ENH] replace no longer compatible scikit-learn make_pipeline with sktime make_pipeline in reducer tests (11155) fkiraly
[ENH] fix compatibility failures in test_reduce (11189) Ayushagrawal-cse
[ENH] remove cross-module imports in classification, regression, transformation, and clustering modules (11213) fkiraly
[ENH] Removal of self._X and self._y from BaseForecaster (10459) Faakhir30
[ENH] remove tests:skip_all from SARIMAX and FreshPRINCE (11233) 8rulerstar
[ENH] better isolation of soft dependencies in TimerS1 vendor module (11224) fkiraly
[ENH] TimesFM 3 multivariate foundation model forecaster (10976) hasanfaesal
[ENH] forecaster capability:update tag and dedicated API tests (11247) fkiraly
[ENH] pretrain support for TimeMoE foundation model (11729) faakhir30
[ENH] add missing license statement to libs.tbats vendor (11157) fkiraly
[ENH] fix tbats incompatibility with latest scikit-learn (11165) fathirramadhan-web
[ENH] add TBATS vendor tests (11170) PiyushKumar74110
[ENH] golden output tests for ChronosForecaster (10692) geetu040
[ENH] golden output tests for Moirai2Forecaster (10687) geetu040
[ENH] golden output tests for Chronos2Forecaster (10677) geetu040
[ENH] golden output tests for KronosForecaster (10997) dongwonmoon
[ENH] golden output tests for FlowStateForecaster (11048) AbdulAliMamnun
[ENH] golden output tests for PatchTSTForecaster (10941) Nischal1425
[ENH] consistent automatic CPU/GPU placement for foundation-model forecasters (11167) geetu040
[ENH] Tafsut foundation model forecaster (11053) aryamanDutta
[ENH] in AutoTS, make silent verbose=-2 parameter the default (11263) fkiraly
[ENH] add NeuralProphet soft dependency bound for pandas<3 (11259) amarjaleelbanbhan
[ENH] clean up soft dependency isolation patterns in networks module (11088) fkiraly
[ENH] Revert #11088 soft dependency isolation as it compromises pickleability, use _safe_import instead (11231) fkiraly
[ENH] Split param_est.stationarity._arch into separate estimator submodules (10552) NAME-ASHWANIYADAV
[ENH] Fix Singular matrix in ImpulseResponseFunction VECM tests (10977) DebojitNath
[ENH] Add second test param sets for ElasticEnsemble, DistFromAligner (10815) NAME-ASHWANIYADAV
[ENH] replace SubLOF in tests with dummy detectors (11152) fkiraly
[ENH] make SubLOF pandas 3 compatible (11136, 11141) fkiraly, TayfurYldz
[ENH] EventTPR, MeanDetectionOffset and FalseAlarmRate for live detection (11197) yash-sangwan
[ENH] remove detection.all based aliasing in detection root imports (11261) fkiraly
[ENH] FreshPRINCE classifier - unskip tests, fix tags (10711) fkiraly
[ENH] Add second test param sets for ElasticEnsemble, DistFromAligner (10815) NAME-ASHWANIYADAV
[ENH] Add second test parameter sets for RandomIntervalClassifier, DummyRegressor, and StackingForecaster (10785) NAME-ASHWANIYADAV
[ENH] Add _get_keras_custom_objects method to base classifier classes for loading deep learning models involving custom layers (10637) srupat
[ENH] Torch-based FCN time series classifier and regressor (9489) kajal-jotwani
[ENH] Fix mutable default arguments in ConvTimeNet backbone (10835) Boubker10
[ENH] add testing tags to deep learning base classes so changes in them trigger tests in child classes (10944) fkiraly
[ENH] torch-based CNTC time-series classifier and regressor (10920) srupat
[ENH] torch based cntc time series classifier and regressor (9397) fnhirwa
[ENH] Migrate WEASEL test skip config from tests._config to estimator tags (10904) JamesBoardman27
[ENH] Move MACNNClassifier test skip config to tags (9814) archittmittal
[ENH] Add doctest examples to deep learning classes (10956) srupat
[ENH] Add second test parameter set for SummaryClassifier, SupervisedTimeSeriesForest (10871) Keykyrios
[ENH] clean up tags for deep learning classifiers and regressors (10953) fkiraly
[ENH] Add second test parameter sets for ContractableBOSS, MUSE, and WEASEL (10934) NAME-ASHWANIYADAV
[ENH] skip ConvTimeNetClassifier test failures on main until resolved (11032) fkiraly
[ENH] Add second test parameter sets for TSFreshClassifier (10999) oberoir080
[ENH] native grid search for TSCGridSearchCV and TSRGridSearchCV (10818) yash-sangwan
[ENH] Make LSTMFCN TensorFlow conv layers dynamic via filter_sizes and kernel_sizes (11078) nicolasdmolina
[ENH] pandas 3 compatibility patches to estimators (11153) yash-sangwan
[ENH] add upper scikit-learn bound to ElasticEnsemble (11154) fkiraly
[ENH] add scikit-learn compatibility bound to TimeSeriesForestClassifier, remove as test case (11176) fkiraly
[ENH] register TimeSeriesForestClassifier tests in tag (11179) fkiraly
[ENH] replace TimeSeriesForestClassifier in more second party tests (11182) fkiraly
[ENH] remove scikit-learn inheritance from TimeSeriesForestClassifier and TimeSeriesForestRegressor (11178) InnoxCodes
[ENH] further scikit-learn 1.8.X compatibility fixes (11185) fkiraly
[ENH] skip sktime in scikit-learn compatibility tests (11186) fkiraly
[ENH] add scikit-learn<1.8 bounds to pyts estimators (11210) fkiraly
[ENH] remove cross-module imports in classification, regression, transformation, and clustering modules (11213) fkiraly
[ENH] remove tests:skip_all from SARIMAX and FreshPRINCE (11233) 8rulerstar
[ENH] ResNet network migration from tf to torch (10762) srupat
[ENH] Add _get_keras_custom_objects method to base classifier classes for loading deep learning models involving custom layers (10637) srupat
[ENH] Hierarchical/Agglomerative Clustering for Time Series (9187) Muhammad-Rebaal
[ENH] Add second test parameter set for TimeSeriesKMedoids (10870) Harshavardhan-28
[ENH] remove cross-module imports in classification, regression, transformation, and clustering modules (11213) fkiraly
[ENH] Add second test parameter sets for RandomIntervalClassifier, DummyRegressor, and StackingForecaster (10785) NAME-ASHWANIYADAV
[BUG] Add _get_keras_custom_objects method to base classifier classes for loading deep learning models involving custom layers (10637) srupat
[ENH] restore test skips in CntcRegressor (10875) fkiraly
[ENH] Torch-based FCN time series classifier and regressor (9489) kajal-jotwani
[ENH] add testing tags to deep learning base classes so changes in them trigger tests in child classes (10944) fkiraly
[ENH] torch-based CNTC time-series classifier and regressor (10920) srupat
[ENH] torch based cntc time series classifier and regressor (9397) fnhirwa
[ENH] Add doctest examples to deep learning classes (10956) srupat
[ENH] clean up tags for deep learning classifiers and regressors (10953) fkiraly
[ENH] native grid search for TSCGridSearchCV and TSRGridSearchCV (10818) yash-sangwan
[ENH] Make LSTMFCN TensorFlow conv layers dynamic via filter_sizes and kernel_sizes (11078) nicolasdmolina
[ENH] remove scikit-learn inheritance from TimeSeriesForestClassifier and TimeSeriesForestRegressor (11178) InnoxCodes
[ENH] remove cross-module imports in classification, regression, transformation, and clustering modules (11213) fkiraly
[ENH] ResNet network migration from tf to torch (10762) srupat
[ENH] ClaSPTransformer second test parameter set (8868) Udayan853
[ENH] address numpy.trapz deprecation and fix test tags in related estimators (10727) fkiraly
[ENH] Add second test parameter set for ClearSky (10788) coding-cosmos
[ENH] Migrate skip-inverse-transform tag to _BaseTag class (10811) DebojitNath
[ENH] Add second FeatureSelection test parameter set (10232) snoopuppy582
[ENH] Add SavitzkyGolayTransformer, HilbertTransformer, and WaveletPacketTransformer (10810) ved197338
[ENH] make FourierFeatures transformation compatible with pandas 3 frequency aliases (10895) yash-sangwan
[ENH] make Imputer mean/median compatible with pandas 3 (10924) yash-sangwan
[ENH] Add unconditional HolidayFeatures test parameters (11045) Sharon-study
[ENH] make TimeSince tests pandas 3 compatible (11134) fkiraly
[ENH] pandas 3 compatibility patches to estimators (11153) yash-sangwan
[ENH] add scikit-learn<1.8 bounds to pyts estimators (11210) fkiraly
[ENH] remove cross-module imports in classification, regression, transformation, and clustering modules (11213) fkiraly
[ENH] Add second test parameter set for DilationMappingTransformer (10657) OfficialAbhinavSingh
[ENH] Add second test parameter sets for PAAlegacy, SAXlegacy, and Hidalgo (10732) NAME-ASHWANIYADAV
[ENH] Refactor test framework step 1: behaviour-preserving fixture rename (#10647) (10661)Faareh-Ahmed, yash-sangwan
[ENH] Refactor test framework step 2: PackageConfig mixin and attribute alignment (#10647) (10807) yash-sangwan
[ENH] Refactor test framework step 3: inherit fixture engine from skbase (#10647) (10862) yash-sangwan
[ENH] Refactor test framework step 4: Inherit TestAllObjects and remove redundant test bodies (11079) yash-sangwan
[ENH] add test that all classes have docstrings and at least one doctest example (10776) fkiraly
[ENH] remove legacy EXCLUDE_ESTIMATORS mechanism from tests._config (11208) fkiraly
[DOC] fix links for pytorch-forecasting (10689) CloseChoice
[DOC] remove ARM troubleshooting section which is no longer applicable (10694) fkiraly
[DOC] fix minor formatting issue in the mlflow integration docstrings (10777) fkiraly
[DOC] add doctest example for DegreeDayFeatures (10817) chintam-dhanush
[DOC] Add doctest example for TimeSeriesDBSCAN (10801) Nikunjsaini07
[DOC] add usage example for Id transformer (10787) baremetaldevx86
[DOC] Add doctest example for Catch22Wrapper #10795 (10799) Nikunjsaini07
[DOC] add Python 3.14 to supported versions in get started guide (10830) AH64-dll
[DOC] add doctest examples for five classes (10796) baremetaldevx86
[DOC] fix DontUpdate docstring (10834) fkiraly
[DOC] add doctest examples for SlidingWindowSegmenter (10836) chintam-dhanush
[DOC] complete VAR docstring for method, missing, freq, dates (10698) shivamlalakiya
[DOC] fix undocumented configs in various foundation models (10630) Faakhir30
[DOC] Fix broken links and typos in ConditionalDeseasonalizer docs (10864) CloseChoice
[DOC] Fix Outdated Dockerfile Example and Table in continuous_integration.rst (10877) JamesBoardman27
[DOC] fix broken links (10865) CloseChoice
[DOC] adds docstring example to PCATransformer (10869) pruthvipatill
[DOC] add doctest example for DirectedChamfer (#10782) (10858) Pragati5-DEBUG
[DOC] add doctest examples for DerivativeSlopeTransformer (10827) chintam-dhanush
[DOC] Move network parameter docs to class docstrings (LSTMFCN, InceptionTime) (9521) sabasiddique1
[DOC] Update links to contributor emoji key (10884) JamesBoardman27
[DOC] Fix FABBA docstring parameter names to match the actual signature (10942) godarrenw
[DOC] add doctest example for CutoffFhSplitter (10939) Valentino-source-dev
[DOC] Add missing randomness tags to forecasting template (10889) Siddhartha23i
[DOC] document property:randomness and capability:random_state in extension templates (10908) chala2001
[DOC] document estimator serialization format (10596) dannymaaz
[DOC] Update governance guidelines link in bug reporting documentation (10913) JamesBoardman27
[DOC] add doctest example for FittedParamExtractor (10963) OfficialAbhinavSingh
[DOC] fix capability:categorical_in_X tag description and incorrect tag defaults in tag registry (10854) siddharth7113
[DOC] add doctest example for SqueezeHierarchy (11013) chrisprown902-wq
[DOC] add doctest example for IgnoreX (11034) oberoir080
[DOC] Fix broken links in vendored fracdiff README (11052) blackorange-mega
[DOC] Add doctest example to SAXlegacy (11066) adan-shahid
[DOC] add doctest example for DistFromAligner (11090) LouisDeconinck
[DOC] add doctest example for MatrixProfileFeatures (11109) shubham5080
[DOC] update README in sktime.libs (11160) Ayushagrawal-cse
[DOC] add doctest example for SignatureKernel (11130) shubham5080
[DOC] add doctest examples for dummy catalogues (11077) HuzaifaAbdulRehman
[DOC] Add usage example to BoxCoxBiasAdjustedForecaster (9429) YadavAkash96
[DOC] add doctest example for SupervisedIntervals (11205) zelihaguven
[DOC] add docstring example to RBFForecaster (11235) fkiraly
[DOC] add doctest examples for TimeBinAggregate (10826) chintam-dhanush
[DOC] add doctest example for WindowSegmenter (10863) chintam-dhanush
[DOC] add capability:multivariate and capability:pairwise to docs (10880) AanchalGupta1162
[DOC] add docstring example for AutoTS forecaster (11262) fkiraly
[DOC] add doctest examples for detection cost functions (10982) jvsch
[MNT] Add CodeQL Analysis Workflow (8406) fkiraly, andoriyaprashant
[MNT] add name to release tag check step for readability (10708) MGPOCKY
[MNT] all-contributors update (10700) github-actions[bot]
[MNT] version upgrade for upper bounds of deps: numpy<2.6, scikit-base<1.2.0, numba<0.67, and some soft deps (10723) fkiraly
[MNT] deprecate the v1 benchmarking framework (#10464) (10636) yash-sangwan
[MNT] all-contributors update (10820) github-actions[bot]
[MNT] refactor test for compatibility with historical dependencies to matrix, add 2025 (10823) fkiraly
[MNT] fix minor bugs in the test.yml CI file (10824) fkiraly
[MNT] Update numba requirement from <0.67,>=0.53 to >=0.53,<0.68 (10860) dependabot[bot]
[MNT] all-contributors update (10878) github-actions[bot]
[MNT] remove extraneous material from package wheel (10892) fkiraly
[MNT] Trigger website docs on stable releases (10638) gthay
[MNT] all-contributors update (10946) github-actions[bot]
[MNT] remove unnecessary from future imports for defunct python 3.9 and lower versions (11033) fkiraly
[MNT] all-contributors update (11042) github-actions[bot]
[MNT] Dummy __init__.py to avoid test fails in the CI (11075) jgyasu
[MNT] Ensure skpro>=2 in CI (11081) fkiraly
[MNT] fix skpro>=2 in notebook dependencies (11084) PiyushKumar74110
[MNT] raise joblib bound to joblib<1.7 (11096) fkiraly
[MNT] add historical 2026 snapshot for testing (11114) fkiraly
[MNT] core dependency upper bound monitor (11095) fkiraly
[MNT] Update skpro requirement from <2.15.0,>=2 to >=2,<2.16.0 (11132) dependabot[bot]
[MNT] do not cancel test-historical-deps all runs if one run fails (11135) fkiraly
[MNT] allow pandas 3 (9279) fkiraly
[MNT] remove upper bounds from notebooks depset (11158) fkiraly
[MNT] Update pytest-randomly requirement from <4.2,>=3.15 to >=3.15,<5.1 (11174) dependabot[bot]
[MNT] raise scikit-learn bound to scikit-learn<1.10.0 (10726) fkiraly
[MNT] update statsmodels upper bounds in pyproject.toml to statsmodels<0.16 (11112) fkiraly
[MNT] all-contributors update (11094) github-actions[bot]
[MNT] all-contributors update (11199) github-actions[bot]
[MNT] Bump github/codeql-action from 4.37.3 to 4.38.1 (10747, 10775, 10849, 10890, 11133, 11241) dependabot[bot]
[MNT] Preserve MPS wrapper metadata for PyTorch 2.14 (11047) PiyushKumar74110
[BUG] Fix unraised exceptions, minor logic and docstring issues (9373) Ashish-Kumar-Dash
[BUG] ignore scikit-learn / sktime conftest modules in scikit-learn estimator registry lookup (11004) geetu040
[BUG] in Pipeline, fix Step._fetch_input_data leaking transformer_names across edge groups (pandas 3) (11142) felipebridge
[BUG] make BaseObject.__eq__ check the class, not just the params (10781) yash-sangwan
[BUG] mean_squared_log_error: remove leftover debug print and stack trace (10816) AlejandroCoronadoN
[BUG] Pin scipy version for neuralprophet (10738) Faakhir30
[BUG] CiscoTSMForecaster predict_proba uses HistogramQPD (10554) Saswatsusmoy
[BUG] Fix CiscoTSMForecaster._predict_proba inconsistency using HistogramQPD (10559) AbiramiR-27
[BUG] Fix mutable default arguments in vendored LagLlama code (10733) Ridadata
[BUG] Fix SCINetForecaster test failure in test_predict_series_name_preserved (10289) dhairya-motta
[BUG] Fix unraised exceptions, minor logic and docstring issues (9373) Ashish-Kumar-Dash
[BUG] set capability:categorical_in_X to False for forecasters whose backend requires numeric X (10855) siddharth7113
[BUG] Fix NaiveVariance with forecasters raising NotImplementedError, such as ExponentialSmoothing on statsmodels 0.15+ (10958) aryamanDutta
[BUG] exclude non-finite scores from AutoResearchForecaster blueprint selection (10965) OfficialAbhinavSingh
[BUG] Fix random_state handling in statsmodels adapter (10972) aryamanDutta
[BUG] fix VARMAX and DynamicFactor failing with integer index on statsmodels 0.15 (11119) Belagum
[BUG] fix Chronos-Bolt long prediction length (11172) Ayushagrawal-cse
[BUG] Greykite API conformance fixes (10755) Faakhir30
[BUG] Fix SquaringResiduals breaking on multi-step fh (10740) Faakhir30
[BUG] fix ARIMA and AutoARIMA out-of-sample prediction with statsmodels 0.15 for unsupported indices (11236) mccool1010
[BUG] Mutable default argument in WindowWarp for LagLLama (10097) direkkakkar319-ops
[BUG] Disable capability:non_contiguous_X for SARIMAX (10741) Faakhir30
[BUG] Fix mutable defaults in LagLlama estimator (10879) tarundb2005-create
[BUG] Preserve pretrained PatchTSMixer configuration (10921) dhruvb2028
[BUG] fix off-by-one in PatchTST prediction_length derived from fh (10940) Nischal1425
[BUG] fix incorrect pd.Series _predict return name in _PytorchForecastingAdapter (11253) fkiraly
[BUG] fix B006 mutable default args in convtimenet backbones (10730) WAHIB-EL-KHADIRI
[BUG] Fix mutable default arguments in ConvTimeNet backbone (10835) Boubker10
[BUG] in DistFromAligner, correct symmetric tag (11110) fkiraly
[BUG] ensure DistFromAligner treats symmetric property directly (11108) webzuweb
[BUG] Resolve bugs in the TapNet Tensorflow implementation (9447) srupat
[BUG] fix deepcopy of estimators using SeqSelfAttentionTorch (10929) srupat
[BUG] fix optimizer passed as an instance not being bound to the network in BaseDeepClassifierPytorch (11035) srupat
[BUG] remove moved ensemble classifiers from classification.compose.__all__ (11104) Anai-Guo
[BUG] fix ConvTimeNetClassifier deepcopy and pickling failures (11099) adity1raut
[BUG] fix typo in DistFromAligner._transform causing TypeError when X2 is None (10841) NAME-ASHWANIYADAV
[BUG] fix missing boolean negation in test_multioutput for TSR (10809)atikulmunna, fkiraly
[BUG] Resolve bugs in the TapNet Tensorflow implementation (9447) srupat
[BUG] fix deepcopy of estimators using SeqSelfAttentionTorch (10929) srupat
[BUG] fix (batch, 1) output handling in BaseDeepRegressorTorch (10932) srupat
[BUG] fix ResNetRegressor broken __init__ (10949) fkiraly
[BUG] Fix optimizer_kwargs silently ignored in BaseDeepRegressorTorch due to wrong guard condition (10926) srupat
[BUG] reenable tests and fix tags for TSFreshFeatureExtractor (10715) fkiraly
[BUG] fix off-by-one in TimeBinAggregate bin_mid index (#10173) (10214) jbbqqf
[BUG] fix "bin_mid" index in TimeBinAggregate in integer case, add test coverage (10175)fkiraly, Tharun-10Dragneel
[BUG] Fix: raise ValueError with correct message in RandomSamplesAugmenter (use f-string) (11026) PiyushKumar74110
[BUG] fix PeakTimeFeature on panel and hierarchical input (10919) Chetansahney
8rulerstar, AanchalGupta1162, AbdulAliMamnun, AbiramiR-27, adan-shahid, adity1raut, AH64-dll, AlejandroCoronadoN, AlvaroBalbin, amarjaleelbanbhan, Anai-Guo, andoriyaprashant, anupamkr1708, archittmittal, aryamanDutta, Ashish-Kumar-Dash, Ayushagrawal-cse, baremetaldevx86, Belagum, blackorange-mega, Boubker10, chala2001, Chetansahney, chintam-dhanush, chrisprown902-wq, CloseChoice, coding-cosmos, dannymaaz, DebD-max, DebojitNath, dhairya-motta, dhruvb2028, direkkakkar319-ops, dongwonmoon, ericjb, Faakhir30, fathirramadhan-web, felipebridge, fkiraly, fnhirwa, gabrielemidulla, geetu040, godarrenw, goyaladitya05, gthay, Harshavardhan-28, hasanfaesal, hdimer, HuzaifaAbdulRehman, InnoxCodes, JamesBoardman27, jbbqqf, jgyasu, jvsch, kajal-jotwani, Keykyrios, KingLizard1020, LouisDeconinck, LunarScoop, luziyi123448-gif, maarcosrmz, marrov, mccool1010, Me-Priyank, MGPOCKY, Muhammad-Rebaal, NAME-ASHWANIYADAV, nicolasdmolina, Nikunjsaini07, Nischal1425, oberoir080, OfficialAbhinavSingh, Partharsid, PiyushKumar74110, Pragati5-DEBUG, pruthvipatill, Ridadata, RobKuebler, sabasiddique1, Saswatsusmoy, Sharon-study, shivamlalakiya, shubham5080, siddharth7113, Siddhartha23i, snoopuppy582, Spicy-source, srupat, sudo-muneeb, tarundb2005-create, TayfurYldz, Tharun-10Dragneel, Udayan853, Valentino-source-dev, ved197338, vedantag17, WAHIB-EL-KHADIRI, webzuweb, wunianze666-netizen, YadavAkash96, yash-sangwan, zelihaguven
Maintenance release with planned changes and deprecations.
Maintenance release with planned changes and deprecations.
Users should note completed deprecations and consult notes below.
Please see our changelog for a description of all changes.
The deprecated ignores-exogeneous-X tag for forecasters has been removed. It has been replaced by capability:exogenous. The logic of the tag is flipped, i.e., if capability:exogenous=True, the estimator can handle exogenous variables.
The deprecated univariate-only tag present in multiple estimator types has been removed. It has been renamed to capability:multivariate, with boolean flip. If capability:multivariate=True, the estimator can handle multivariate data.
Users and maintainers of third party estimators should update their code and estimators to use the new tags.
The deprecated scitype:y tag present in forecasters has been removed. It is renamed to capability:multivariate. If capability:multivariate=True, the estimator can handle multivariate data, if False, it cannot. Values of the old tag map as follows onto the new tag: "univariate" maps onto False, "multivariate" and "both" map onto True. Users and maintainers of third party estimators should update their code and estimators to use the new tag.
The deprecated legacy API for global forecasting (y in predict) has been removed. Users should use the new global forecasting API through pretrain. The capability:global_forecasting tag is deprecated and will be removed entirely in version 1.2.0; until then, access will raise a warning. Users should use the capability:pretrain tag instead. Tag queries will not be redirected due to simultaneous presence of both tags in versions prior.
Minor release with scheduled deprecations and changes.
Users should note completed deprecations and consult notes below.
The deprecated ignores-exogeneous-X tag for forecasters has been removed. It has been replaced by capability:exogenous . The logic of the tag is flipped, i.e., if capability:exogenous=True , the estimator can handle exogenous variables. Users and maintainers of third party estimators should update their code and estimators to use the new tag.
The deprecated univariate-only tag present in multiple estimator types has been removed. It has been renamed to capability:multivariate , with boolean flip. If capability:multivariate=True , the estimator can handle multivariate data. Users and maintainers of third party estimators should update their code and estimators to use the new tag.
forecaster and transformation capability tags for exogenous variables and multivariate data, capability:exogenous and capability:multivariate , replace the ignores-exogeneous-X and univariate-only tags, respectively. Values False and True map to True and False , respectively.
The deprecated scitype:y tag present in forecasters has been removed. It is renamed to capability:multivariate . If capability:multivariate=True , the estimator can handle multivariate data, if False , it cannot. Values of the old tag map as follows onto the new tag: "univariate" maps onto False , "multivariate" and "both" map onto True .
The deprecated legacy API for global forecasting ( y in predict ) has been removed. Users should use the new global forecasting API through pretrain . The capability:global_forecasting tag is deprecated and will be removed entirely in version 1.2.0; until then, access will raise a warning. Users should use the capability:pretrain tag instead. Tag queries will not be redirected due to simultaneous presence of both tags in versions prior.
Feature release. Please see our changelog for a description of all changes.
Feature release. Please see our changelog for a description of all changes.
multiple foundation models for forecasting added:
Arps Decline Curve Analysis forecasters (#10258) @scuervo91
EvoForestTSWM transformer (frozen closed-form feature extractor) (#10458) @kayuksel
time series regression model_evaluation and RegressionBenchmark (#10409) @NAME-ASHWANIYADAV
results analyzers for v2 benchmarking (#10482) @yash-sangwan
MiniRocketMultivariateCython, a Cython (numba-free) implementation of MiniRocketMultivariate with no JIT warmup, faster transform, and optional thread parallelism via n_jobs (#10425) @sssilvar
HyperTreeNetARForecaster from hypertrees-forecasting (#10544) @aminehd
Chen & Yang (2004) forecast accuracy metrics (#9364) @michaelellis003
@Abelarm, @adan-shahid, @adrynalean, @AMBRA7592, @aminehd, @Ankit-1204, @archittmittal, @AYUSH27112021, @CloseChoice, @Faakhir30, @fkiraly, @geetu040, @gnanadeep256, @harish885, @jgyasu, @joshdunnlime, @julian-fong, @kayuksel, @kerimkarakan, @Kevin23-design, @kiwoongyoon, @Krishna21435, @kumarshobhit, @loulanyue, @michaelellis003, @NAME-ASHWANIYADAV, @narges-aibi, @Nischal1425, @OfficialAbhinavSingh, @onkar717, @phoeenniixx, @pyarchana, @Rishav23av, @SAY-5, @scuervo91, @Si-ra-kri, @siddharth7113, @SimonBlanke, @snoopuppy582, @Solaris-star, @sssilvar, @stephanielees, @ubermensch19, @Vbhatt03, @vedantag17, @vedhakoushik, @VenkateshHJoshi, @vortex-wq, @wali-reheman, @XAheli, @yash-sangwan, @ziad-ashraf7
Full Changelog: https://github.com/sktime/sktime/compare/v1.0.1...v1.0.2
multiple foundation models for forecasting added:
Aurora multimodal foundation model forecaster ( #10417 ) @Faakhir30
Cisco TSFM forecaster ( #10444 ) @vedantag17
Falcon-X forecaster ( #10430 ) @vedantag17
MIRA medical time series foundation model ( #10398 ) @Faakhir30
Moirai 2.0 foundation model forecasters ( #9678 , #10333 ) @ubermensch19 , @vedantag17
Thuml Sundial forecasting foundation model ( #10428 ) @geetu040
Toto-2.0 forecasting foundation model ( #10385 ) @siddharth7113
WindFM forecasting foundation model ( #10384 ) @geetu040
Arps Decline Curve Analysis forecasters ( #10258 ) @scuervo91
EvoForestTSWM transformer (frozen closed-form feature extractor) ( #10458 ) @kayuksel
time series regression model_evaluation and RegressionBenchmark ( #10409 ) @NAME-ASHWANIYADAV
results analyzers for v2 benchmarking ( #10482 ) @yash-sangwan
MiniRocketMultivariateCython , a Cython (numba-free) implementation of MiniRocketMultivariate with no JIT warmup, faster transform, and optional thread parallelism via n_jobs ( #10425 ) @sssilvar
HyperTreeNetARForecaster from hypertrees-forecasting ( #10544 ) @aminehd
Chen & Yang (2004) forecast accuracy metrics ( #9364 ) @michaelellis003
[ENH] point duplicative sktime.dependencies imports to scikit-base ( #10405 ) @fkiraly
[ENH] exclude sktime.libs from estimator discovery ( #10401 ) @Nischal1425
[ENH] remove cross-module imports - part 1 ( #10512 ) @fkiraly
[ENH] replace generic raise Exception with specific ValueError in 9 locations ( #10514 ) @vedhakoushik
[ENH] Add estimator checks for deepcopy support ( #10497 ) @pyarchana
[ENH] Fault-tolerant benchmark runs ( #10461 ) @jgyasu
[ENH] Crash-safe result storage and resuming for benchmarking experiments ( #10297 ) @jgyasu
[ENH] results analzyers for v2 benchmarking (strategy pattern) ( #10482 ) @yash-sangwan
[ENH] temporarily skip test_forecastingbenchmark_global_mode until fixed ( #10570 ) @fkiraly
[ENH] regression model_evaluation and RegressionBenchmark ( #10409 ) @NAME-ASHWANIYADAV
[ENH] remove benchmarking.test_base prior to deprecation of old framework ( #10655 ) @fkiraly
[ENH] Chen & Yang (2004) forecast accuracy metrics (#8681) ( #9364 ) @michaelellis003
[ENH] centralize supported distance-string registry in distances ( #9555 ) @ziad-ashraf7
[ENH] Moirai 2.0 foundation model forecasters ( #9678 ) @ubermensch19
[ENH] Moirai 2 foundation model forecasters ( #10333 ) @vedantag17
[ENH] WindFM forecasting foundation model ( #10384 ) @geetu040
[ENH] TimeLLM: remove trust_remote_code and other fixes ( #10328 ) @geetu040
[ENH] update TinyTimeMixerForecaster with new model support, pretraining, and compatibility improvements ( #10395 ) @geetu040
[ENH] preserve pretrain state during reset, without override of public methods ( #10410 ) @SimonBlanke
[ENH] remove cross-module imports - part 2 ( #10526 ) @fkiraly
[ENH] Toto-2.0 forecasting foundation model (Toto2Forecaster) ( #10385 ) @siddharth7113
[ENH] Aurora multimodal foundation model forecaster ( #10417 ) @Faakhir30
[ENH] Cisco TSFM forecaster ( #10444 ) @vedantag17
[ENH] MIRA medical time series foundation model for forecasting by Microsoft ( #10398 ) @Faakhir30
[ENH] Momentfm caching and lazy load addition ( #10494 ) @vedantag17
[ENH] rename timesfm2_forecaster module to timesfm2 ( #10539 ) @fkiraly
[ENH] minor improvements to DummyGlobalForecaster code ( #10562 ) @fkiraly
[ENH] in test_forecastingbenchmark_global_mode , replace soft dependent estimator with DummyGlobalForecaster ( #10561 ) @fkiraly
[ENH] GreyKiteForecaster - clean up tags ( #10571 ) @fkiraly
[ENH] Arps Decline Curve Analysis forecasters ( #10258 ) @scuervo91
[ENH] Falcon-X Forecaster ( #10430 ) @vedantag17
[ENH] interface to hypertrees-forecasting : HyperTreeNetARForecaster ( #10544 ) @aminehd
[ENH] add _predict_proba to Arps DCA forecasters via Empirical distribution ( #10611 ) @Nischal1425
[ENH] second test parameter sets for OptionalPassthrough and DontUpdate (#3429) ( #9816 ) @archittmittal
[ENH] expose exogenous variable support in TotoForecaster ( #10421 ) @siddharth7113
[ENH] FlowState predict_proba using HistogramQPD ( #10558 ) @loulanyue
[ENH] TimerS1Forecaster : consistent predict_proba via HistogramQPD ( #10577 ) @Solaris-star
[ENH] Moved SquaringResiduals skipped test from tests._config to estimator tags ( #8907 ) @Abelarm
[ENH] temporarily skip faulty part of test_pred_int_tag ( #9034 ) @fkiraly
[ENH] Thuml Sundial forecasting foundation model ( #10428 ) @geetu040
[ENH] Deprecate global forecasting in flowstate forecaster ( #10424 ) @Faakhir30
[ENH] Split param_est.seasonality module into one-estimator-per-file submodules ( #10509 ) @fkiraly
[ENH] Move ARLagOrderSelector test skip config to estimator tags ( #10443 ) @vortex-wq
[ENH] remove cross-module imports - part 3 ( #10524 ) @NAME-ASHWANIYADAV
[ENH] Move ClusterSegmenter test skip rule to estimator tag ( #10199 ) @Kevin23-design
[ENH] Improve plotting change points ( #9831 ) @stephanielees
[ENH] add tests and proper documentation for refactored activation inputs in TSC and TSR ( #10362 ) @Faakhir30
[ENH] TSPulse caching addition ( #10507 ) @vedantag17
[ENH] Added second test parameter to ColumnEnsembleClassifier ( #7685 ) @AYUSH27112021
[ENH] Move MLPRegressor test skip config from tests._config to estimator tags ( #10438 ) @gnanadeep256
[ENH] add tests and proper documentation for refactored activation inputs in TSC and TSR ( #10362 ) @Faakhir30
[ENH] regression model_evaluation and RegressionBenchmark ( #10409 ) @NAME-ASHWANIYADAV
[ENH] Add tests for dilation mapping transformer ( #10423 ) @kumarshobhit
[ENH] MiniRocketMultivariateCython , a numba-free Cython MiniRocket transform ( #10425 ) @sssilvar
[ENH] EvoForestTSWM transformer (frozen closed-form feature extractor) ( #10458 ) @kayuksel
[ENH] Add second test parameter set for PCATransformer ( #10492 ) @narges-aibi
[ENH] Add second test parameter set for PAA ( #10489 ) @narges-aibi
[ENH] add Test params to shapelet transform ( #10470 ) @julian-fong
[ENH] Add second PAA test parameter set ( #10231 ) @snoopuppy582
[ENH] Add second test parameter set for PlateauFinder ( #10490 ) @narges-aibi
[ENH] Add second test parameter set for SlopeTransformer ( #10575 ) @NAME-ASHWANIYADAV
[ENH] second test parameter sets for OptionalPassthrough and DontUpdate (#3429) ( #9816 ) @archittmittal
[ENH] Add second test parameter set for MovingWindow ( #9396 ) @Vbhatt03
[ENH] Add second test parameter set for PaddingTransformer ( #10656 ) @OfficialAbhinavSingh
[ENH] add pytest check for cross-module imports between type-specific modules ( #10479 ) @Nischal1425
[ENH] tests:specific tag and CI and run estimator-specific pytest modules from TestAllObjects in virtual machines ( #10649 ) @fkiraly
[ENH] Add some tests:specific tags for VM-tested estimators with dedicated test modules ( #10651 ) @Copilot
[ENH] Weekly spot test for forecasters integrity ( #10365 ) @jgyasu
[ENH] Weekly spot test for classifiers integrity ( #10367 ) @jgyasu
[ENH] Weekly spot test for transformers integrity ( #10368 ) @jgyasu
[DOC] Fix typo: forecasters horizon → forecasting horizon ( #10406 ) @wali-reheman
[DOC] fix a number of broken links in API reference ( #10436 ) @fkiraly
[DOC] fix make_forecasting_scorer func docstring ( #9382 ) @adan-shahid
[DOC] clarify make_forecasting_scorer signature ( #10435 ) @adrynalean
[DOC] fix incorrect data type in PanelGluontsPandas docstring ( #10408 ) @harish885
[DOC] fix minor typo in AcorrLjungbox docstring ( #10473 ) @fkiraly
[DOC] update developer documentation for estimators with cython dependency ( #10471 ) @fkiraly
[DOC] update soft dependency management guide for developers to 1.0 patterns ( #10472 ) @fkiraly
[DOC] add missing TiRexForeaster to API reference ( #10505 ) @fkiraly
[DOC] fix docstrings of post_init ( #10506 ) @fkiraly
[DOC] fix broken contributors hall of fame badge in README ( #10547 ) @Nischal1425
[DOC] Document HierarchicalPolarsEager mtype ( #10318 ) @AMBRA7592
[DOC] fix typo in fracdiff README (imcorporate → incorporate) ( #9424 ) @Si-ra-kri
[DOC] document the cov parameter in BaseForecaster.predict_var ( #10587 ) @SAY-5
[DOC] fix stale no-update_params description in add_task and evaluate ( #10593 ) @Nischal1425
[DOC] fix formatting in augmenter docstrings ( #10643 ) @fkiraly
[DOC] fix minor typos in tag registry ( #10653 ) @fkiraly
[DOC] Add docstring examples to statsforecast forecasters ( #9568 ) @Krishna21435
[DOC] Add missing import to WeightedEnsembleClassifier docstring example ( #9453 ) @VenkateshHJoshi
[DOC] Fix docstring typos and formatting in igts.py and boxcox.py ( #10079 ) @onkar717
[DOC] improve CNNClassifier parameter docstring formatting ( #9427 ) @Si-ra-kri
[DOC] Add XAheli (Aheli Poddar) to all-contributors ( #10521 ) @XAheli
[DOC] Fix docstrings and typos in benchmarking module ( #10501 ) @NAME-ASHWANIYADAV
[DOC] Add Examples to StatsForecast adapter docstrings ( #9567 ) @Krishna21435
[MNT] Fix pytest failures due to skip with fixture removal in 9.0 release ( #10407 ) @jgyasu
[MNT] Update pytest requirement from <9.1,>=7.4 to >=7.4,<9.2 ( #10414 ) @dependabot[bot]
[MNT] testing for R based estimators ( #8909 ):user: ericjb , @fkiraly
[MNT] Bump actions/checkout from 6 to 7 ( #10431 ) @dependabot[bot]
[MNT] lint deprecated _contrib module ( #10441 ) @fkiraly
[MNT] all-contributors update ( #10499 ) @github-actions[bot]
[MNT] Update skpro requirement from <2.14.0,>=2 to >=2,<2.15.0 ( #10513 ) @dependabot[bot]
[MNT] Update lightning requirement from <2.6,>=2.0 to >=2.0,<2.7 ( #10257 ) @dependabot[bot]
[MNT] Update lightning requirement from <2.6,>=2.0 to >=2.0,<2.7 ( #10525 ) @dependabot[bot]
[MNT] all-contributors update ( #10527 ) @github-actions[bot]
[MNT] isolate torch in huggingface-hub PyTorchModelHubMixin imports ( #10530 ) @fkiraly
[MNT] Bump actions/setup-node from 6 to 7 ( #10546 ) @dependabot[bot]
[MNT] add numba dependency to notebooks test depset ( #10553 ) @fkiraly
[MNT] Remove dorny 3rd party GHA and dead GHA workflows ( #10556 ) @fkiraly
[MNT] all-contributors update ( #10567 ) @github-actions[bot]
[MNT] Remove redundant test causing sporadic data download failures ( #10565 ) @jgyasu
[MNT] Bump actions/setup-python from 6 to 7 ( #10595 ) @dependabot[bot]
[MNT] add skpro to VM test depsets, remove skchange from binder depset ( #10639 ) @fkiraly
[MNT] all-contributors update ( #10646 ) @github-actions[bot]
[MNT] all-contributors update ( #10437 ) @github-actions[bot]
[MNT] Enable HF_TOKEN in CI ( #10465 ) @geetu040
[BUG] Correct benchmark.add() ’s inconsistent API ( #10463 ) @jgyasu
[BUG] Fix missing f-string in _SktimeRegistry.register warning ( #10467 ) @NAME-ASHWANIYADAV
[BUG] Change per-forecast-step metric computation to aggregate over horizon in OWA ( #10548 ) @jgyasu
[BUG] Fix BaseResults.save missing NotImplementedError raise ( #10171 ) @Rishav23av
[BUG] Fix invalid LTSFLinearForecaster spec in test_craft ( #10515 ) @kumarshobhit
[BUG] Fix wrong converter registered for pd-long to nested_univ Panel conversion ( #10181 ) @onkar717
[BUG] fix pytorch-forecasting adapter predict with exogenous X ( #10383 ) @Nischal1425
[BUG] Fix test errors for pytorch-forecasting estimators ( #10415 ) @CloseChoice
[BUG] Solve the multiple failures in pytorch-forecasting forecasters ( #10412 ) @Nischal1425 , @phoeenniixx , @CloseChoice
[BUG] fix parameter-dependent tags discarded before super().init() in foundation model forecasters ( #10528 ) @kerimkarakan
[BUG] Preserve pretrained state across repeated fit calls ( #10529 ) @SimonBlanke
[BUG] preliminary fix for incorrect setting of _pretrained_attrs in case init already sets fitted attrs ( #10535 ) @fkiraly
[BUG] in ARCH , fix constructor tag setting: move set_tags to after super call and add tests ( #10613 ) @joshdunnlime
[BUG] fix context_length lookup for use_source_package Chronos-Bolt ( #10594 ) @Nischal1425
[BUG] Align TinyTimeMixer padding masks with Granite-TSFM ( #10627 ) @geetu040
[BUG] Guard MOIRAIForecaster vendored imports lazily ( #10495 ) @kiwoongyoon
[BUG] fix SCINetForecaster pretrain test failures via smaller test params ( #10496 ) @Nischal1425
[BUG] Fix KNeighborsTimeSeriesClassifier.neighbor failures ( #8093 ) @Ankit-1204
[BUG] Replace bitwise & with logical and in boolean validation logic ( #10180 ) @onkar717
@Abelarm , @adan-shahid , @adrynalean , @AMBRA7592 , @aminehd , @Ankit-1204 , @archittmittal , @AYUSH27112021 , @CloseChoice , @Faakhir30 , @fkiraly , @geetu040 , @gnanadeep256 , @harish885 , @jgyasu , @joshdunnlime , @julian-fong , @kayuksel , @kerimkarakan , @Kevin23-design , @kiwoongyoon , @Krishna21435 , @kumarshobhit , @loulanyue , @michaelellis003 , @NAME-ASHWANIYADAV , @narges-aibi , @Nischal1425 , @OfficialAbhinavSingh , @onkar717 , @phoeenniixx , @pyarchana , @Rishav23av , @SAY-5 , @scuervo91 , @Si-ra-kri , @siddharth7113 , @SimonBlanke , @snoopuppy582 , @Solaris-star , @sssilvar , @stephanielees , @ubermensch19 , @Vbhatt03 , @vedantag17 , @vedhakoushik , @VenkateshHJoshi , @vortex-wq , @wali-reheman , @XAheli , @yash-sangwan , @ziad-ashraf7
multiple foundation models for forecasting added:
Aurora multimodal foundation model forecaster (10417) Faakhir30
Cisco TSFM forecaster (10444) vedantag17
Falcon-X forecaster (10430) vedantag17
MIRA medical time series foundation model (10398) Faakhir30
Moirai 2.0 foundation model forecasters (9678, 10333) ubermensch19, vedantag17
Thuml Sundial forecasting foundation model (10428) geetu040
Toto-2.0 forecasting foundation model (10385) siddharth7113
WindFM forecasting foundation model (10384) geetu040
Arps Decline Curve Analysis forecasters (10258) scuervo91
EvoForestTSWM transformer (frozen closed-form feature extractor) (10458) kayuksel
time series regression model_evaluation and RegressionBenchmark (10409) NAME-ASHWANIYADAV
results analyzers for v2 benchmarking (10482) yash-sangwan
MiniRocketMultivariateCython, a Cython (numba-free) implementation of MiniRocketMultivariate with no JIT warmup, faster transform, and optional thread parallelism via n_jobs (10425) sssilvar
HyperTreeNetARForecaster from hypertrees-forecasting (10544) aminehd
Chen & Yang (2004) forecast accuracy metrics (9364) michaelellis003
[ENH] point duplicative sktime.dependencies imports to scikit-base (10405) fkiraly
[ENH] exclude sktime.libs from estimator discovery (10401) Nischal1425
[ENH] remove cross-module imports - part 1 (10512) fkiraly
[ENH] replace generic raise Exception with specific ValueError in 9 locations (10514) vedhakoushik
[ENH] Add estimator checks for deepcopy support (10497) pyarchana
[ENH] Fault-tolerant benchmark runs (10461) jgyasu
[ENH] Crash-safe result storage and resuming for benchmarking experiments (10297) jgyasu
[ENH] results analzyers for v2 benchmarking (strategy pattern) (10482) yash-sangwan
[ENH] temporarily skip test_forecastingbenchmark_global_mode until fixed (10570) fkiraly
[ENH] regression model_evaluation and RegressionBenchmark (10409) NAME-ASHWANIYADAV
[ENH] remove benchmarking.test_base prior to deprecation of old framework (10655) fkiraly
[ENH] Chen & Yang (2004) forecast accuracy metrics (#8681) (9364) michaelellis003
[ENH] centralize supported distance-string registry in distances (9555) ziad-ashraf7
[ENH] Moirai 2.0 foundation model forecasters (9678) ubermensch19
[ENH] Moirai 2 foundation model forecasters (10333) vedantag17
[ENH] WindFM forecasting foundation model (10384) geetu040
[ENH] TimeLLM: remove trust_remote_code and other fixes (10328) geetu040
[ENH] update TinyTimeMixerForecaster with new model support, pretraining, and compatibility improvements (10395) geetu040
[ENH] preserve pretrain state during reset, without override of public methods (10410) SimonBlanke
[ENH] remove cross-module imports - part 2 (10526) fkiraly
[ENH] Toto-2.0 forecasting foundation model (Toto2Forecaster) (10385) siddharth7113
[ENH] Aurora multimodal foundation model forecaster (10417) Faakhir30
[ENH] Cisco TSFM forecaster (10444) vedantag17
[ENH] MIRA medical time series foundation model for forecasting by Microsoft (10398) Faakhir30
[ENH] Momentfm caching and lazy load addition (10494) vedantag17
[ENH] rename timesfm2_forecaster module to timesfm2 (10539) fkiraly
[ENH] minor improvements to DummyGlobalForecaster code (10562) fkiraly
[ENH] in test_forecastingbenchmark_global_mode, replace soft dependent estimator with DummyGlobalForecaster (10561) fkiraly
[ENH] GreyKiteForecaster - clean up tags (10571) fkiraly
[ENH] Arps Decline Curve Analysis forecasters (10258) scuervo91
[ENH] Falcon-X Forecaster (10430) vedantag17
[ENH] interface to hypertrees-forecasting: HyperTreeNetARForecaster (10544) aminehd
[ENH] add _predict_proba to Arps DCA forecasters via Empirical distribution (10611) Nischal1425
[ENH] second test parameter sets for OptionalPassthrough and DontUpdate (#3429) (9816) archittmittal
[ENH] expose exogenous variable support in TotoForecaster (10421) siddharth7113
[ENH] FlowState predict_proba using HistogramQPD (10558) loulanyue
[ENH] TimerS1Forecaster: consistent predict_proba via HistogramQPD (10577) Solaris-star
[ENH] Moved SquaringResiduals skipped test from tests._config to estimator tags (8907) Abelarm
[ENH] temporarily skip faulty part of test_pred_int_tag (9034) fkiraly
[ENH] Thuml Sundial forecasting foundation model (10428) geetu040
[ENH] Deprecate global forecasting in flowstate forecaster (10424) Faakhir30
[ENH] Split param_est.seasonality module into one-estimator-per-file submodules (10509) fkiraly
[ENH] Move ARLagOrderSelector test skip config to estimator tags (10443) vortex-wq
[ENH] remove cross-module imports - part 3 (10524) NAME-ASHWANIYADAV
[ENH] Move ClusterSegmenter test skip rule to estimator tag (10199) Kevin23-design
[ENH] Improve plotting change points (9831) stephanielees
[ENH] add tests and proper documentation for refactored activation inputs in TSC and TSR (10362) Faakhir30
[ENH] TSPulse caching addition (10507) vedantag17
[ENH] Added second test parameter to ColumnEnsembleClassifier (7685) AYUSH27112021
[ENH] Move MLPRegressor test skip config from tests._config to estimator tags (10438) gnanadeep256
[ENH] add tests and proper documentation for refactored activation inputs in TSC and TSR (10362) Faakhir30
[ENH] regression model_evaluation and RegressionBenchmark (10409) NAME-ASHWANIYADAV
[ENH] Add tests for dilation mapping transformer (10423) kumarshobhit
[ENH] MiniRocketMultivariateCython, a numba-free Cython MiniRocket transform (10425) sssilvar
[ENH] EvoForestTSWM transformer (frozen closed-form feature extractor) (10458) kayuksel
[ENH] Add second test parameter set for PCATransformer (10492) narges-aibi
[ENH] Add second test parameter set for PAA (10489) narges-aibi
[ENH] add Test params to shapelet transform (10470) julian-fong
[ENH] Add second PAA test parameter set (10231) snoopuppy582
[ENH] Add second test parameter set for PlateauFinder (10490) narges-aibi
[ENH] Add second test parameter set for SlopeTransformer (10575) NAME-ASHWANIYADAV
[ENH] second test parameter sets for OptionalPassthrough and DontUpdate (#3429) (9816) archittmittal
[ENH] Add second test parameter set for MovingWindow (9396) Vbhatt03
[ENH] Add second test parameter set for PaddingTransformer (10656) OfficialAbhinavSingh
[ENH] add pytest check for cross-module imports between type-specific modules (10479) Nischal1425
[ENH] tests:specific tag and CI and run estimator-specific pytest modules from TestAllObjects in virtual machines (10649) fkiraly
[ENH] Add some tests:specific tags for VM-tested estimators with dedicated test modules (10651) Copilot
[ENH] Weekly spot test for forecasters integrity (10365) jgyasu
[ENH] Weekly spot test for classifiers integrity (10367) jgyasu
[ENH] Weekly spot test for transformers integrity (10368) jgyasu
[DOC] Fix typo: forecasters horizon → forecasting horizon (10406) wali-reheman
[DOC] fix a number of broken links in API reference (10436) fkiraly
[DOC] fix make_forecasting_scorer func docstring (9382) adan-shahid
[DOC] clarify make_forecasting_scorer signature (10435) adrynalean
[DOC] fix incorrect data type in PanelGluontsPandas docstring (10408) harish885
[DOC] fix minor typo in AcorrLjungbox docstring (10473) fkiraly
[DOC] update developer documentation for estimators with cython dependency (10471) fkiraly
[DOC] update soft dependency management guide for developers to 1.0 patterns (10472) fkiraly
[DOC] add missing TiRexForeaster to API reference (10505) fkiraly
[DOC] fix docstrings of __post_init__ (10506) fkiraly
[DOC] fix broken contributors hall of fame badge in README (10547) Nischal1425
[DOC] Document HierarchicalPolarsEager mtype (10318) AMBRA7592
[DOC] fix typo in fracdiff README (imcorporate → incorporate) (9424) Si-ra-kri
[DOC] document the cov parameter in BaseForecaster.predict_var (10587) SAY-5
[DOC] fix stale no-update_params description in add_task and evaluate (10593) Nischal1425
[DOC] fix formatting in augmenter docstrings (10643) fkiraly
[DOC] fix minor typos in tag registry (10653) fkiraly
[DOC] Add docstring examples to statsforecast forecasters (9568) Krishna21435
[DOC] Add missing import to WeightedEnsembleClassifier docstring example (9453) VenkateshHJoshi
[DOC] Fix docstring typos and formatting in igts.py and boxcox.py (10079) onkar717
[DOC] improve CNNClassifier parameter docstring formatting (9427) Si-ra-kri
[DOC] Add XAheli (Aheli Poddar) to all-contributors (10521) XAheli
[DOC] Fix docstrings and typos in benchmarking module (10501) NAME-ASHWANIYADAV
[DOC] Add Examples to StatsForecast adapter docstrings (9567) Krishna21435
[MNT] Fix pytest failures due to skip with fixture removal in 9.0 release (10407) jgyasu
[MNT] Update pytest requirement from <9.1,>=7.4 to >=7.4,<9.2 (10414) dependabot[bot]
[MNT] testing for R based estimators (8909)ericjb, fkiraly
[MNT] Bump actions/checkout from 6 to 7 (10431) dependabot[bot]
[MNT] lint deprecated _contrib module (10441) fkiraly
[MNT] all-contributors update (10499) github-actions[bot]
[MNT] Update skpro requirement from <2.14.0,>=2 to >=2,<2.15.0 (10513) dependabot[bot]
[MNT] Update lightning requirement from <2.6,>=2.0 to >=2.0,<2.7 (10257) dependabot[bot]
[MNT] Update lightning requirement from <2.6,>=2.0 to >=2.0,<2.7 (10525) dependabot[bot]
[MNT] all-contributors update (10527) github-actions[bot]
[MNT] isolate torch in huggingface-hub PyTorchModelHubMixin imports (10530) fkiraly
[MNT] Bump actions/setup-node from 6 to 7 (10546) dependabot[bot]
[MNT] add numba dependency to notebooks test depset (10553) fkiraly
[MNT] Remove dorny 3rd party GHA and dead GHA workflows (10556) fkiraly
[MNT] all-contributors update (10567) github-actions[bot]
[MNT] Remove redundant test causing sporadic data download failures (10565) jgyasu
[MNT] Bump actions/setup-python from 6 to 7 (10595) dependabot[bot]
[MNT] add skpro to VM test depsets, remove skchange from binder depset (10639) fkiraly
[MNT] all-contributors update (10646) github-actions[bot]
[MNT] all-contributors update (10437) github-actions[bot]
[MNT] Enable HF_TOKEN in CI (10465) geetu040
[BUG] Correct benchmark.add()'s inconsistent API (10463) jgyasu
[BUG] Fix missing f-string in _SktimeRegistry.register warning (10467) NAME-ASHWANIYADAV
[BUG] Change per-forecast-step metric computation to aggregate over horizon in OWA (10548) jgyasu
[BUG] Fix BaseResults.save missing NotImplementedError raise (10171) Rishav23av
[BUG] Fix invalid LTSFLinearForecaster spec in test_craft (10515) kumarshobhit
[BUG] Fix wrong converter registered for pd-long to nested_univ Panel conversion (10181) onkar717
[BUG] fix pytorch-forecasting adapter predict with exogenous X (10383) Nischal1425
[BUG] Fix test errors for pytorch-forecasting estimators (10415) CloseChoice
[BUG] Solve the multiple failures in pytorch-forecasting forecasters (10412) Nischal1425, phoeenniixx, CloseChoice
[BUG] fix parameter-dependent tags discarded before super().__init__() in foundation model forecasters (10528) kerimkarakan
[BUG] Preserve pretrained state across repeated fit calls (10529) SimonBlanke
[BUG] preliminary fix for incorrect setting of _pretrained_attrs in case __init__ already sets fitted attrs (10535) fkiraly
[BUG] in ARCH, fix constructor tag setting: move set_tags to after super call and add tests (10613) joshdunnlime
[BUG] fix context_length lookup for use_source_package Chronos-Bolt (10594) Nischal1425
[BUG] Align TinyTimeMixer padding masks with Granite-TSFM (10627) geetu040
[BUG] Guard MOIRAIForecaster vendored imports lazily (10495) kiwoongyoon
[BUG] fix SCINetForecaster pretrain test failures via smaller test params (10496) Nischal1425
[BUG] Fix KNeighborsTimeSeriesClassifier.neighbor failures (8093) Ankit-1204
[BUG] Replace bitwise & with logical and in boolean validation logic (10180) onkar717
Abelarm, adan-shahid, adrynalean, AMBRA7592, aminehd, Ankit-1204, archittmittal, AYUSH27112021, CloseChoice, Faakhir30, fkiraly, geetu040, gnanadeep256, harish885, jgyasu, joshdunnlime, julian-fong, kayuksel, kerimkarakan, Kevin23-design, kiwoongyoon, Krishna21435, kumarshobhit, loulanyue, michaelellis003, NAME-ASHWANIYADAV, narges-aibi, Nischal1425, OfficialAbhinavSingh, onkar717, phoeenniixx, pyarchana, Rishav23av, SAY-5, scuervo91, Si-ra-kri, siddharth7113, SimonBlanke, snoopuppy582, Solaris-star, sssilvar, stephanielees, ubermensch19, Vbhatt03, vedantag17, vedhakoushik, VenkateshHJoshi, vortex-wq, wali-reheman, XAheli, yash-sangwan, ziad-ashraf7
Mini-release fixing readthedocs documentation build errors.
Mini-release fixing readthedocs documentation build errors.
Major release with some breaking changes.
Major release with some breaking changes.
Please see our changelog for a description of all changes.
@12jaspreetsingh, @Abelarm, @abhimanyudalal1, @adan-shahid, @adrynalean, @algojogacor, @alphaleporus, @Aniketsy, @archittmittal, @Ask-812, @AyushI7G, @benHeid, @Chetansahney, @CloseChoice, @crocmons, @DCchoudhury15, @Deep-Axe, @direkkakkar319-ops, @EmanAbdelhaleem, @Faakhir30, @fkiraly, @FlyingDragon112, @G26karthik, @Gautam-Bharadwaj, @geetu040, @Giggitycountless, @goyaladitya05, @gun29may, @gupta-tilak, @haosenwang1018, @Ironankit525, @JATAYU000, @jdhruv555, @jgyasu, @junaidaslam2006, @Kevin23-design, @kpal002, @marrov, @Mohit25f101, @neha222222, @neharoy3, @NestroyMusoke, @nikol-damyanova, @Nischal1425, @nvphungdev, @OldPatrick, @onkar717, @paramsureliya, @PewterZz, @piyushbiraje, @PragnyaKhandelwal, @pranavvp16, @priyanshuharshbodhi1, @purvanshjoshi, @R2-STAR, @rakshaak29, @RecreationalMath, @Rusheel86, @sabasiddique1, @SajeelHussain, @Saloni-0465, @shaun0927, @SimonBlanke, @Sonika-19, @Spandan-Mishra, @SpitFire19, @srupat, @TenFinges, @thecaptain789, @thisisrick25, @varun-kht, @vedantag17, @xenonnn4w, @yash-sangwan, @zstinnett3
Full Changelog: https://github.com/sktime/sktime/compare/v0.40.1...v1.0.0
Major release with some breaking API improvements. Users and developers should consult the changelog below for API changes.
pretraining and global forecasting API for forecasters ( #9056 ) @SimonBlanke
AutoResearchForecaster for agentic, data-driven improvement of forecasting pipelines ( #9855 ) @benHeid
a large selection of additional state-of-art foundation models is now supported: Chronos2, FalconTST, Flowstate, Kronos, Lagllama, Mantis, MomentFM variants, PatchTSMixer, Timer, TSpulse, Bytedance Timer-S1, TimesFM 2
API consistency and memory usage for foundation models has been substantially improved; compatible dependency ranges have been widened for many forecasters
Time series classifiers and regressors have been migrated to full torch native implementations, @Faakhir30 and others
Benchmark catalogues, M-Forecasting Competition and TSC bakeoff Catalogues ( #9246 , #9138 ) @jgyasu
Johansen cointegration and impulse-response functions for ARIMA class models ( #9040 , #9217 ) @OldPatrick
tsfeatures and tsfel feature extraction transformers ( #9082 , #9107 ) @Faakhir30
DegreeDayFeatures transformer for heating and cooling degree days ( #9231 ) @zstinnett3
sktime now requires scikit-base 1.0 or higher.
The dependency set all_extras has been substantially reduced. This no longer contains any estimator specific optional dependencies, only optional framework dependencies. Estimator specific dependencies are tracked in the python_dependencies tag, and should be installed from there on a by-estimator basis.
numpy (core dependency) bounds have been updated to <2.5
pandas (core dependency) bounds have been updated to <3.0
scipy (core dependency) bounds have been updated to <2.0
the extension template for all estimators has been refined by adding methods dynamic_tags and post_init . The constructor init is now reserved for setting init args to self only. Dynamic tags should be set in dynamic_tags only, and any other light initialization logic should be carried out in post_init or at the start of _fit . For more details, extenders should consult the updated extension templates.
new object type: "catalogue" , with base class BaseCatalogue . Catalogues contain collections of estimators, metrics, datasets, and/or any other type of object. They can be natively passed to the sktime benchmarking framework to set up pre-defined benchmarks with ease, e.g., estimators from the M4 forecasting competition.
passing an integer fh to forecasters is now interpreted as range(1, fh+1) . For example, passing fh=3 will forecast the next 3 periods, and is equivalent to passing fh=[1, 2, 3] , or fh=range(1, 4) .
forecasters now possess a programmatic API point for pre-training, the pretrain method. Global forecasters, indicated by the capability:pretrain tag, can be pre-trained via pretrain on global training data on which later no forecast is queried via predict . All forecasters have pretrain , but it is a no-op for those with capability:pretrain tag value False .
a new tag capability:non_contiguous_X has been introduced. This boolean tag describes whether a forecaster can handle non-contiguous future-known exogenous data X in predict .
The sktime.libs.pykalman module has been removed in favour of the pykalman package, also maintained by GC.OS and the sktime team. Users of pykalman based estimators in sktime can continue to use them, and the estimator API itself will not change. However, users will need to install pykalman separately, as it is no longer included in sktime .
The ignores-exogeneous-X tag for forecasters is replaced by capability:exogenous . The logic of the tag is flipped, i.e., if capability:exogenous=True , the estimator can handle exogenous variables. Users and maintainers of third party estimators should update their code and estimators to use the new tag.
the univariate-only tag present in multiple estimator types is renamed to capability:multivariate , with boolean flip. If capability:multivariate=True , the estimator can handle multivariate data. Users and maintainers of third party estimators should update their code and estimators to use the new tag.
forecaster and transformation capability tags for exogenous variables and multivariate data, capability:exogenous and capability:multivariate , replace the ignores-exogeneous-X and univariate-only tags, respectively. Values False and True map to True and False , respectively. The old tags are still aliased until 1.1.0, and a warning will be raised when accessing the old tags. From 1.1.0 onwards, an exception will be raised.
Deprecated tags python_dependencies_alias and univariate-metric have now been removed.
the scitype:y tag present in forecasters is renamed to capability:multivariate . If capability:multivariate=True , the estimator can handle multivariate data, if False , it cannot. Values of the old tag map as follows onto the new tag: "univariate" maps onto False , "multivariate" and "both" map onto True . Until 1.1.0, the old tag, scitype:y , can still be used, but will emit a warning. From 1.1.0 onwards, an exception will be raised. Users and maintainers of third party estimators should update their code and estimators to use the new tag.
the deprecated tag capability:pred_var has been removed.
the old estimator type series-annotator for specific detectors has been removed after deprecation period, and replaced by the umbrella tag detector .
all sktime estimators from the skchange package have been merged into sktime.detection proper, after cessation of maintenance by skchange as part of a separate package.
detection and transformations module folder structure has been flattened, most estimators are now directly in a submodule of sktime.detection or sktime.transformations . In particular, the modules sktime.transformations.series and sktime.transformations.panel no longer exist, instead users should import sktime.transformations.series.[module] directly from sktime.transformations.[module] . The old import paths will still continue to work for until the last 1.X version.
[ENH] catalogues in sktime ( #8749 ) @jgyasu
[ENH] make _safe_import robust against package import errors ( #9193 ) @fkiraly
[ENH] craft deserialization utility can now also index scikit-learn estimators ( #9228 ) @fkiraly
[ENH] Removed series-as-features module ( #9519 ) @RecreationalMath
[ENH] Clean up exogenous capability tag references ( #9254 ) @junaidaslam2006
[ENH] change dependency error in estimator init to warning ( #9962 ) @fkiraly
[ENH] move dynamic tag setting to dedicated dynamic_tags method ( #10140 ) @fkiraly
[ENH] remove old deprecated tag python_dependencies_alias ( #10298 ) @fkiraly
[ENH] Add _check_estimator_deps to base classes missing soft dependency enforcement ( #9805 ) @Deep-Axe
[ENH] MeanSquaredLogarithmicError (MSLE) metric ( #9140 ) @alphaleporus
[ENH] add extension templates for forecasting metrics ( #9844 ) @NestroyMusoke
[ENH] Allow list , and dict of estimators in benchmark.add ( #9916 ) @jgyasu
[ENH] Overall Weighted Average (OWA) Score ( #9111 ) @jgyasu
[ENH] moved PresplitFilesCV and SingleSplit to split module ( #9519 ) @RecreationalMath
[ENH] Fix constructor dependency handling in catalogues ( #10295 ) @fkiraly
[ENH] remove old deprecated tag univariate-metric ( #10298 ) @fkiraly
[ENH] M4 Forecasting Competition Catalogue ( #9246 ) @jgyasu
[ENH] TSC Bake Off 2017 Catalogue ( #9138 ) @jgyasu
[ENH] Prevent dataset downloads in catalogue tests ( #10324 ) @jgyasu
[ENH] Refactor BaseCatalogue ( #10320 ) @jgyasu
[ENH] Fix capability:exogenous tags for FM forecasters (#8523) ( #9163 ) @jdhruv555
[ENH] capability:non_contiguous_X tag and skip non-contiguous X tests for affected estimators ( #9091 ) @EmanAbdelhaleem
[ENH] reduce forecaster coupling to self._y attribute - part 1 ( #9195 ) @fkiraly
[ENH] ensure test data in test_update_with_exogenous is nonsingular ( #9204 ) @fkiraly
[ENH] support for neuralforecast new versions ( #8750 ) @thisisrick25
[ENH] Enable prediction intervals for AutoTS ( #9203 ) @TenFinges
[ENH] remove “only multivariate” forecaster property ( #8343 ) @fkiraly
[ENH] add missing capability:non_contiguous_X tags ( #9213 ) @fkiraly
[ENH] Enable exogenous support for AutoTS ( #9210 ) @TenFinges
[ENH] Make ForecastingOptunaSearchCV tests compatible without CmaEsSampler ( #9269 ) @junaidaslam2006
[ENH] TransformedTargetForecaster.predict_proba , via transformed distribution (#9287) ( #9346 ) @xenonnn4w
[ENH] pretraining and global forecasting API for forecasters ( #9056 ) @SimonBlanke
[ENH] xLSTM forecaster ( #8575 ) @vedantag17
[ENH] add pretraining feature for pytorch-based forecasters ( #9541 ) @SimonBlanke
[ENH] deprecate forecaster scitype:y tag in favour of capability:multivariate ( #9212 ) @fkiraly
[ENH] skip StatsForecastMSTL testing on negative data with multiplicative seasonality ( #9807 ) @fkiraly
[ENH] in forecasting tutorial, make pandas calls upwards compatible ( #9820 ) @fkiraly
[ENH] Add second test param sets for Prophet forecaster ( #8855 ) @piyushbiraje
[ENH] Chronos 2 foundation model forecaster ( #9823 ) @priyanshuharshbodhi1
[ENH] update DummyGlobalForecaster tags to 1.0 ( #9868 ) @fkiraly
[ENH] Lagllama forecaster ( #9194 ) @pranavvp16
[ENH] fix optuna sampler fixtures in tests ( #9771 ) @fkiraly
[ENH] Move StackingForecaster test skip from _config to estimator tags ( #9975 ) @paramsureliya
[ENH] Move SquaringResiduals test skip rule to estimator tag ( #9866 ) @Kevin23-design
[ENH] simplify IgnoreX init logic ( #10094 ) @fkiraly
[ENH] remove deprecated cmaes sampler from optuna facing tests ( #9216 ) @fkiraly
[ENH] TimerForecaster for zero-shot time series forecasting ( #9744 ) @PewterZz
[ENH] use BaseForecaster direct inheritance for foundation model forecasters ( #9276 ) @SimonBlanke
[ENH] move more forecaster code to post_init and dynamic_tags . ( #10146 ) @fkiraly
[ENH] reduce forecaster coupling to self._y attribute - part 2 ( #9202 ) @fkiraly
[ENH] Allow FMs to handle unequal length hierarchical data ( #10125 ) @jgyasu
[ENH] Timer: vendor architecture code into sktime.libs and fix deps ( #10327 ) @geetu040
[ENH] declare statsmodels as soft dep of HCrystalBallAdapter ( #10104 ) @kpal002
[ENH] add tests:vm tag to HCrystalBallAdapter ( #10243 ) @fkiraly
[ENH] Unequal length TS test and tag for forecasters ( #10078 ) @jgyasu
[ENH] pandas 3 compatibility fixes ( #9821 ) @fkiraly
[ENH] mantis forecaster ( #10266 ) @vedantag17
[ENH] Add params count in _fit for _NeuralForecastAdapter ( #8428 ) @Faakhir30
[ENH] Interface TSPulse for classification ( #10278 ) @Faakhir30
[ENH] remove old deprecated tags univariate-metric ( #10298 ) @fkiraly
[ENH] neuralprophet forecaster ( #8348 ) @vedantag17
[ENH] uniformize forecasting tuner parameters and docstrings ( #10247 ) @fkiraly
[ENH] TimesFM 2.x forecasting foundation models ( #10291 ) @geetu040
[ENH] Bytedance Timer-S1 forecasting foundation model ( #10315 ) @geetu040
[ENH] temporarily skip failing test for BoxCoxBiasAdjustedForecaster ( #10338 ) @fkiraly
[ENH] remove _forecaster suffixes in forecasting submodules ( #10337 ) @fkiraly
[ENH] Update HFTransformersForecaster tags and docstring ( #10355 ) @fkiraly
[ENH] fix deprecated import in test_hf_transformers_forecaster.py ( #10369 ) @fkiraly
[ENH] AutoResearchForecaster - agentic improvement of forecaster blueprint ( #9855 ) @benHeid
[ENH] FalconTST forecasting foundation model ( #10361 ) @geetu040
[ENH] Kronos forecasting foundation model ( #10347 ) @geetu040
[ENH] deduplicate ARIMA test skips ( #10372 ) @fkiraly
[ENH] Flowstate forecasting foundation model by IBM ( #10329 ) @Faakhir30
[ENH] PatchTSMixer Forecaster by IBM ( #10330 ) @Faakhir30
[ENH] passing fh: int to forecast range(1, fh + 1) ( #10378 ) @fkiraly
[ENH] Johansen cointegration ( #9040 ) @OldPatrick
[ENH] Impulse Response Function/Parameters for Uni/Multivariate Time-Series ( #9217 ) @OldPatrick
[ENH] in AlignerDTW and AlignerDTWfromDist , add check for incorrect dtw package installation ( #9214 ) @DCchoudhury15
[ENH] momentfm time series anomaly detection foundation model ( #9500 ) @Faakhir30
[ENH] remove redundant code in BaseDetector.init ( #9963 ) @fkiraly
[ENH] Offline univariate Hampel filter detector ( #9888 ) @SpitFire19
[ENH] remove deprecated scipy Kulsinski distance from tests ( #10191 , #10252 ) @direkkakkar319-ops
[ENH] Natively merge skchange core detection algorithms (#9636) ( #9719 ) @yash-sangwan
[ENH] add change conditional test skips to new tests in detection module ( #10359 ) @fkiraly
[ENH] tspulse for anomaly detection ( #10308 ) @Faakhir30
[ENH] Hidden layer droput uniformization in the deep learning models ( #9109 ) @abhimanyudalal1
[ENH] Expose dropout parameters in CNTCNetwork and TapNetNetwork ( #9208 ) @xenonnn4w
[ENH] torch based MCDCNN regressor and classifier ( #9232 ) @Faakhir30
[ENH] torch based MLP classifier ( #9042 ) @RecreationalMath
[ENH] torch based LSTMFCN regressor and classifier ( #9298 ) @Faakhir30
[ENH] torch based InceptionTime regressor and classifier#9305 ( #9314 ) @Faakhir30
[ENH] Implement torch based tapnet time series classifier and regressor ( #9412 ) @srupat
[ENH] Make conv layers dynamic in MCDCNNClassifierTorch and MCDCNNRegressorTorch ( #9342 ) @Faakhir30
[ENH] Make MLP TensorFlow layers dynamic via n_layers and hidden_dim params ( #9607 ) @varun-kht
[ENH] Make FCN TensorFlow layers dynamic via n_layers , filter_sizes , and kernel_sizes ( #9618 ) @varun-kht
[ENH] Implement ConvTran deep learning estimator for classification and regression in PyTorch ( #9550 ) @srupat
[ENH] replace 1 bare except clauses with except Exception ( #9502 ) @haosenwang1018
[ENH] torch based InceptionTime regressor and classifier#9305 ( #9314 ) @Faakhir30
[ENH] Remove X.shape checking, single labeled data warnings from TSC and TSR end estimators. ( #9450 ) @Faakhir30
[ENH] Refactor instantiation and validation of activations in torch models ( #9622 ) @Faakhir30
[ENH] in torch based TSC and TSR, add support for evaluation metrics from torchmetrics ( #9390 ) @Faakhir30
[ENH] Move SimpleRNNClassifier test skip config to tags ( #9559 ) @Spandan-Mishra
[ENH] Move MCDCNNClassifier test skip config to estimator tags ( #9813 ) @archittmittal
[ENH] torch based MACNN regressor and classifier ( #9305 ) @Faakhir30
[ENH] torch based CNN regressor and classifier ( #9330 ) @Faakhir30
[ENH] mantis classifier foundational model ( #10282 ) @vedantag17
[ENH] TSC Bake Off 2017 Catalogue ( #9138 ) @jgyasu
[ENH] in torch based TSC and TSR, add support for evaluation metrics from torchmetrics ( #9390 ) @Faakhir30
[ENH] Complete stale 0.41.0 deprecations for CNN/RNN activation defaults ( #10207 ) @onkar717
[ENH] add second parameter set for TimeSeriesKernelKMeans ( #9960 ) @fkiraly
[ENH] torch based RNN regressor ( #9013 ) @RecreationalMath
[ENH] Hidden layer droput uniformization in the deep learning models ( #9109 ) @abhimanyudalal1
[ENH] Expose dropout parameters in CNTCNetwork and TapNetNetwork ( #9208 ) @xenonnn4w
[ENH] torch based MCDCNN regressor and classifier ( #9232 ) @Faakhir30
[ENH] torch based LSTMFCN regressor and classifier ( #9298 ) @Faakhir30
[ENH] torch based InceptionTime regressor and classifier#9305 ( #9314 ) @Faakhir30
[ENH] torch based MLP regressor ( #9223 ) @Faakhir30
[ENH] Make conv layers dynamic in MCDCNNClassifierTorch and MCDCNNRegressorTorch ( #9342 ) @Faakhir30
[ENH] Make MLP TensorFlow layers dynamic via n_layers and hidden_dim params ( #9607 ) @varun-kht
[ENH] Make FCN TensorFlow layers dynamic via n_layers , filter_sizes , and kernel_sizes ( #9618 ) @varun-kht
[ENH] Implement ConvTran deep learning estimator for classification and regression in PyTorch ( #9550 ) @srupat
[ENH] replace 1 bare except clauses with except Exception ( #9502 ) @haosenwang1018
[ENH] torch based InceptionTime regressor and classifier#9305 ( #9314 ) @Faakhir30
[ENH] Remove X.shape checking, single labeled data warnings from TSC and TSR end estimators. ( #9450 ) @Faakhir30
[ENH] Refactor instantiation and validation of activations in torch models ( #9622 ) @Faakhir30
[ENH] in torch based TSC and TSR, add support for evaluation metrics from torchmetrics ( #9390 ) @Faakhir30
[ENH] Implement torch based tapnet time series classifier and regressor ( #9412 ) @srupat
[ENH] torch based MACNN regressor and classifier ( #9305 ) @Faakhir30
[ENH] torch based CNN regressor and classifier ( #9330 ) @Faakhir30
[ENH] in torch based TSC and TSR, add support for evaluation metrics from torchmetrics ( #9390 ) @Faakhir30
[ENH] Complete stale 0.41.0 deprecations for CNN/RNN activation defaults ( #10207 ) @onkar717
[ENH] Allow ColumnwiseTransformer to work directly with sklearn transformers using TabularToSeriesAdaptor ( #9132 ) @crocmons
[ENH] tsfel feature extraction transformer ( #9082 ) @Faakhir30
[ENH] tsfeatures feature extraction transformer ( #9107 ) @Faakhir30
[ENH] DegreeDayFeatures transformer for heating and cooling degree days ( #9231 ) @zstinnett3
[ENH] Move RandomIntervalSegmenter test skip config to tags ( #9313 ) @Mohit25f101
[ENH] In WindowSummarizer , replaced deprecated pandas grouper with keys ( #9367 ):user: marrov , @adan-shahid
[ENH] Remove deprecated ColumnTransformer ( #9979 ) @fkiraly
[ENH] remove stale RowTransformer entries from test config ( #10129 ) @Giggitycountless
[ENH] remove upper numpy bound in MatrixProfileTransformer and address segmentation faults ( #10339 ) @CloseChoice
[ENH] Flatten transformations module ( #10371 , #10168 ) @fkiraly , @CloseChoice
[ENH] estimator tag for extra testing dependencies ( #9207 ) @fkiraly
[ENH] Prevent shell injection in sktime/utils/git_diff.py ( #9616 ) @Ironankit525
[ENH] update test_run_test_for_class to take into account VM-only tests ( #9836 ) @fkiraly
[ENH] update test_run_test_for_class to take into account VM-only tests - part 2 ( #9865 ) @fkiraly
[ENH] Add tests:vm to pytorch estimators base layer ( #9542 ) @Faakhir30
[ENH] simplify show_versions by removing alias handling ( #10356 ) @fkiraly
[ENH] make show_versions normalization proof, e.g., for huggingface-hub ( #10364 ) @fkiraly
[ENH] complete deprecation of fdiff(mode="full") - replace DeprecationWarning with ValueError ( #9662 ) @PewterZz
[DOC] Fix broken pytorch_forecasting documentation references ( #8923 ) @G26karthik
[DOC] callbacks missing in the docstring of some deep learning models ( #9160 ) @Aniketsy
[DOC] add missing callbacks to FCNClassifier docstring ( #9183 ) @12jaspreetsingh
[DOC] correct typo ‘occurence’ to ‘occurrence’ ( #9339 ) @thecaptain789
[DOC] correct typo ‘neccessary’ to ‘necessary’ ( #9349 ) @thecaptain789
[DOC] Typo fixes and docstring cleanup in clustering and distances ( #9670 ) @Gautam-Bharadwaj
[DOC] fix typo: remove duplicate ‘to’ in docstring ( #9683 ) @AyushI7G
[DOC] Programmatic list of estimator types in the documentation ( #9647 ) @purvanshjoshi
[DOC] Added an explanatory note before deprecation examples ( #9061 ) @RecreationalMath
[DOC] Fixed Docs and broken hyperlinks ( #9445 ) @gun29may
[DOC] improve programmatic list of estimator types in API reference ( #9774 ) @neharoy3
[DOC] Fix np.darray typo to np.ndarray in docs and comments ( #9826 ) @Saloni-0465
[DOC] Fix typos in docstrings and comments ( #9673 ) @Ask-812
[DOC] Fix typo in forecasting notebook (choic → choice) ( #9987 ) @R2-STAR
[DOC] improve formatting in BaseForecaster docstrings ( #10047 ) @fkiraly
[DOC] shorten KalmanFilterTransformerSIMD to remove pykalman dependency ( #10057 ) @fkiraly
[DOC] Fix documented default value of error_score parameter in forecasting benchmark ( #10062 ) @benHeid
[DOC] Fix “to to” double-word typo in docstrings ( #10077 ) @direkkakkar319-ops
[DOC] improve DirectReductionForecaster docstring ( #10091 ) @fkiraly
[DOC] improved docstring for craft , deps , and imports serialization utilities ( #9261 ) @fkiraly
[DOC] Skip fit and predict in doc example of Prophetverse and HierarchicalProphet ( #10172 ) @jgyasu
[DOC] clarify post_init_ contract w.r.t. memory and compute heavy operations ( #10188 ) @fkiraly
[DOC] Fix incorrect reference in SlidingWindowSplitter docstring ( #9825 ) @Saloni-0465
[DOC] Document dynamic_tags constructor method in extension templates ( #10296 ) @fkiraly
[DOC] ForecastingRandomizedSearchCV usage example in docstring ( #10294 ) @SajeelHussain
[DOC] ForecastingGridSearchCV Examples indentation fix in docstring ( #10303 ) @SajeelHussain
[DOC] add mentions of python 3.14 in missing documentation locations ( #10317 ) @fkiraly
[DOC] in developer contribution guide, update documentation on pull request stacking ( #10304 ) @fkiraly
[DOC] Fix WDTW docstring typo ( #10331 ) @Saloni-0465
[DOC] clarify TemporalTrainTestSplitter docstring ( #10341 ) @adrynalean
[DOC] Make spelling consistent: “exogeneous” to “exogenous” ( #10135 ) @direkkakkar319-ops
[MNT] add vm estimators to test-all workflow ( #9112 ) @neha222222
Bump actions/download-artifact from 6 to 7 ( #9167 ) @dependabot[bot]
Bump actions/upload-artifact from 5 to 6 ( #9168 ) @dependabot[bot]
[MNT] bound torch<2.6 ( #9171 ) @fkiraly
[MNT] CI step with pinned dependencies in August 2024 ( #9170 ) @fkiraly
[MNT] scipy 1.17 compatibility patch ( #9238 ) @fkiraly
[ENH] dockerfiles: Python 3.13 and 3.14 support ( #9130 ) @crocmons
[MNT] Dependabot: Update sphinx-gallery requirement from <0.20.0 to <0.21.0 ( #9148 ) @dependabot[bot]
[MNT] bound lightning<2.6 ( #9173 ) @fkiraly
Update skpro requirement from <2.11.0,>=2 to >=2,<2.12.0 ( #9179 ) @dependabot[bot]
[MNT] merge release wheels testing workflows ( #9190 ) @fkiraly
[MNT] prioritize empty return in run_test_vm ( #9192 ) @fkiraly
[ENH] Prevent test_all to run on forks ( #9123 ) @Abelarm
[MNT] move autots testing to dedicated VM ( #9206 ) @fkiraly
Update autots requirement from <0.7,>=0.6.1 to >=0.6.1,<1.1 ( #9147 ) @dependabot[bot]
[MNT] Dependabot: Update networkx requirement from <3.5 to <3.7 ( #9221 ) @dependabot[bot]
[MNT] remove pandas 1 CI ( #9387 ) @fkiraly
Bump actions/download-artifact from 7 to 8 ( #9513 ) @dependabot[bot]
Bump actions/upload-artifact from 6 to 7 ( #9514 ) @dependabot[bot]
[MNT] Updated lower bound for skbase version support ( #9723 ) @RecreationalMath
[MNT] all-contributors update ( #9803 ) @github-actions[bot]
[MNT] increase numpy bound to <2.5 ( #9817 ) @fkiraly
[MNT] clean up depsets for 1.0 release ( #9386 ) @fkiraly
[MNT] simplify all_extras soft dependency set ( #9870 ) @fkiraly
[MNT] Use uv date cutoffs instead of extras for old dependency version testing ( #9872 ):user: zanieb , @fkiraly
Update skpro requirement from <2.12.0,>=2 to >=2,<2.13.0 ( #9886 ) @dependabot[bot]
[MNT] Dependabot: Update sphinx-issues requirement from <6.0.0 to <7.0.0 ( #9907 ) @dependabot[bot]
Update scikit-base requirement from <0.14.0,>=0.13.1 to >=0.13.1,<1.1.0 ( #9952 ) @dependabot[bot]
Update pytest-randomly requirement from <4.1,>=3.15 to >=3.15,<4.2 ( #9961 ) @dependabot[bot]
[MNT] make test-est CI element require test-nosoftdeps ( #9974 ) @fkiraly
[MNT] all-contributors update ( #9927 ) @github-actions[bot]
[MNT] Update CODEOWNERS to active core developers ( #9978 ) @fkiraly
[MNT] point pykalman import to package after its restoration ( #9976 ) @fkiraly
[MNT] remove module specific tests from CI ( #10011 ) @fkiraly
[MNT] chunk estimator VM tests to ensure maximum matrix size of 256 is not exceeded in test-est ( #10012 ) @fkiraly
[MNT] Dependabot: Update sphinx-gallery requirement from <0.21.0 to <0.22.0 ( #10032 ) @dependabot[bot]
[MNT] in test-est CI, ensure all dependencies are resolved at once ( #10046 ) @fkiraly
[MNT] all-contributors update ( #10033 ) @github-actions[bot]
[MNT] all-contributors update ( #10149 ) @github-actions[bot]
[MNT] add the artificial-intelligence keyword to pyproject.toml ( #10242 ) @fkiraly
[MNT] Update Makefile install target for pyproject builds ( #10244 ) @nvphungdev
[MNT] replace deprecated setup.py install with pip install in Makefile ( #10254 ) @algojogacor
[MNT] all-contributors update ( #10334 ) @github-actions[bot]
Update skpro requirement from <2.13.0,>=2 to >=2,<2.14.0 ( #10353 ) @dependabot[bot]
[MNT] remove _contrib folder from package ( #10370 ) @fkiraly
[MNT] all-contributors update ( #9182 ) @github-actions[bot]
[MNT] all-contributors update ( #10210 ) @github-actions[bot]
[MNT] Deduplicate sktime.utils.dependencies methods ( #9403 ) @RecreationalMath
[MNT] Update Makefile install target for pyproject builds ( #10244 ) @nvphungdev
[MNT] 1.0.0 deprecations and change actions ( #10377 ) @fkiraly
[BUG] fixes to craft utility and datatypes returns ( #9688 ) @fkiraly
[BUG] Add missing raise keywords to Exceptions across core modules ( #10060 ) @rakshaak29
[BUG] fix cutoff in evaluate global mode, use y_hist not global y_train ( #10373 ) @Nischal1425
[BUG] Fix load_fpp3 by updating CRAN package version strings ( #10147 ) @Rusheel86
[BUG] fixes to craft utility and datatypes returns ( #9688 ) @fkiraly
[BUG] fix gluonts_ListDataset_panel metadata ( #10061 ) @goyaladitya05
[BUG] Fix hf_xet dependency name to hf-xet in moirai_forecaster.py ( #9189 ) @gupta-tilak
[BUG] fix VECM wrong tag and internal variable name handling ( #9205 ) @fkiraly
[BUG] Skip non-converged AutoETS fits during model selection ( #9292 ) @sabasiddique1
[BUG] bound transformers dependency ( #9596 ) @SimonBlanke
[BUG] Resolve Prophet crash by robustly handling stan_backend and fixing test dimensions ( #9234 ) @DCchoudhury15
[BUG] fix API non-conformance in Chronos2Forecaster ( #9924 ) @fkiraly
[BUG] Chronos2 truncation axis fix and added a regression test ( #9955 ) @Chetansahney
[BUG] fix LagLlamaForecaster._predict self-mutation of _is_range_index ( #9921 ) @shaun0927
[BUG] Fix PatchTSTForecaster compatibility with transformers>=5.0 ( #9874 ) @neha222222
[BUG] Fix TinyTimeMixerForecaster AttributeError with transformers>=5.0 ( #9875 ) @neha222222
[BUG] fix GreykiteForecaster test flags ( #10150 ) @fkiraly
[BUG] ChronosForecaster compatibility for transformers 5.X ( #10132 ) @nikol-damyanova
[BUG] Fix Chronos-Bolt incompatibility with transformers >=5.x and restore test coverage ( #10167 ) @Sonika-19
[BUG] Fix error message for forecaster count validation for TransformedTargetForecaster ( #10043 ) @benHeid
[BUG] Fix TimesFM dependency imports and enable VM testing ( #10265 ) @geetu040
[BUG] fix TimesFMForecaster multiton method name ( #10302 ) @fkiraly
[BUG] fix pmdarima forecasters y.name preservation in predict ( #10276 ) @fkiraly
[BUG] add missing tests:libs and tests:vm tags for forecasters with internal dependencies ( #10277 ) @fkiraly
[BUG] fix HFTransformersForecaster compatibility with transformers >= 4.41 ( #10087 ) @Nischal1425
[BUG] fix predict_interval index mismatch for ARIMA / AutoARIMA full in-sample fh ( #10319 ) @kpal002
[BUG] Fix ChronosForecaster T5Stack initialization on transformers>=5.0 ( #10085 ) @sabasiddique1
[ENH] fork necessary functions from seasonal package ( #9155 ) @JATAYU000
[BUG] Fix IndexError in twe_alignment_path default bounding matrix ( #9196 ) @DCchoudhury15
[BUG] Remove duplicate BasetDetector._transform_score function ( #9835 ) @direkkakkar319-ops
[BUG] fix BaseDetector sparse-to-dense conversion for pd.DataFrame input ( #9890 ) @direkkakkar319-ops
[BUG] Fix NaN handling in ProximityForest _stdp calculation ( #9186 ) @TenFinges
[BUG] Improve default hidden layer activation in CNN based classifier and regressor ( #9050 ) @RecreationalMath
[BUG] Fix padding macnn conv layer tensorflow ( #9307 ) @Faakhir30
[BUG] fix optimizer instantiation in BaseDeepClassifierPyTorch class ( #9359 ) @RecreationalMath
[BUG] RocketClassifier - fix expected outputs + readd test (fixes issue #7921) ( #9152 ) @SimonBlanke
[BUG] Followup bug fixing on Inceptiontime ( #9643 ) @Faakhir30
[BUG] Fix TSCOptCV incorrect base class inheritance ( #10098 ) @direkkakkar319-ops
[BUG] Fix parameter order in MCDCNN estimators’ docstring and constructor & add type hints ( #9393 ) @Faakhir30
[BUG] Improve default hidden layer activation in CNN based classifier and regressor ( #9050 ) @RecreationalMath
[BUG] Fix padding macnn conv layer tensorflow ( #9307 ) @Faakhir30
[BUG] Followup bug fixing on Inceptiontime ( #9643 ) @Faakhir30
[BUG] Fix parameter order in MCDCNN estimators’ docstring and constructor & add type hints ( #9393 ) @Faakhir30
[BUG] Fix ColumnwiseTransformer sklearn support using coerce_scitype ( #9200 ) @DCchoudhury15
[BUG] Fix WindowSummarizer bfill across multiindex groups ( #9243 ) @marrov
[BUG] Guard logsignature post-rescaling and add regression test ( #9285 ) @sabasiddique1
[BUG] Replace .rename() with .rename_categories() to prevent FutureWarning in DateTimeFeatures ( #9433 ) @FlyingDragon112
[BUG] Fix TSBootstrap test tag typo tests:vms -> tests:vm ( #9959 ) @PragnyaKhandelwal
[BUG] remove eval from TransformIf condition evaluation ( #9993 ) @PragnyaKhandelwal
[BUG] Fix duplicated mean in std_transformer ( #10186 ) @Saloni-0465
[BUG] fix mlflow test failure - double use of database reference ( #9185 ) @jdhruv555
[BUG] Failing test case in fracdiff ( #10185 ) @direkkakkar319-ops
@12jaspreetsingh , @Abelarm , @abhimanyudalal1 , @adan-shahid , @adrynalean , @algojogacor , @alphaleporus , @Aniketsy , @archittmittal , @Ask-812 , @AyushI7G , @benHeid , @Chetansahney , @CloseChoice , @crocmons , @DCchoudhury15 , @Deep-Axe , @direkkakkar319-ops , @EmanAbdelhaleem , @Faakhir30 , @fkiraly , @FlyingDragon112 , @G26karthik , @Gautam-Bharadwaj , @geetu040 , @Giggitycountless , @goyaladitya05 , @gun29may , @gupta-tilak , @haosenwang1018 , @Ironankit525 , @JATAYU000 , @jdhruv555 , @jgyasu , @junaidaslam2006 , @Kevin23-design , @kpal002 , @marrov , @Mohit25f101 , @neha222222 , @neharoy3 , @NestroyMusoke , @nikol-damyanova , @Nischal1425 , @nvphungdev , @OldPatrick , @onkar717 , @paramsureliya , @PewterZz , @piyushbiraje , @PragnyaKhandelwal , @pranavvp16 , @priyanshuharshbodhi1 , @purvanshjoshi , @R2-STAR , @rakshaak29 , @RecreationalMath , @Rusheel86 , @sabasiddique1 , @SajeelHussain , @Saloni-0465 , @shaun0927 , @SimonBlanke , @Sonika-19 , @Spandan-Mishra , @SpitFire19 , @srupat , @TenFinges , @thecaptain789 , @thisisrick25 , @varun-kht , @vedantag17 , @xenonnn4w , @yash-sangwan , @zstinnett3
Major release with some breaking API improvements. Users and developers should consult the changelog below for API changes.
pretraining and global forecasting API for forecasters (9056) SimonBlanke
AutoResearchForecaster for agentic, data-driven improvement of forecasting pipelines (9855) benHeid
a large selection of additional state-of-art foundation models is now supported: Chronos2, FalconTST, Flowstate, Kronos, Lagllama, Mantis, MomentFM variants, PatchTSMixer, Timer, TSpulse, Bytedance Timer-S1, TimesFM 2
API consistency and memory usage for foundation models has been substantially improved; compatible dependency ranges have been widened for many forecasters
Time series classifiers and regressors have been migrated to full torch native implementations, Faakhir30 and others
Benchmark catalogues, M-Forecasting Competition and TSC bakeoff Catalogues (9246, 9138) jgyasu
Johansen cointegration and impulse-response functions for ARIMA class models (9040, 9217) OldPatrick
tsfeatures and tsfel feature extraction transformers (9082, 9107) Faakhir30
DegreeDayFeatures transformer for heating and cooling degree days (9231) zstinnett3
sktime now requires scikit-base 1.0 or higher.
The dependency set all_extras has been substantially reduced. This no longer contains any estimator specific optional dependencies, only optional framework dependencies. Estimator specific dependencies are tracked in the python_dependencies tag, and should be installed from there on a by-estimator basis.
numpy (core dependency) bounds have been updated to <2.5
pandas (core dependency) bounds have been updated to <3.0
scipy (core dependency) bounds have been updated to <2.0
the extension template for all estimators has been refined by adding methods __dynamic_tags__ and __post_init__. The constructor __init__ is now reserved for setting __init__ args to self only. Dynamic tags should be set in __dynamic_tags__ only, and any other light initialization logic should be carried out in __post_init__ or at the start of _fit. For more details, extenders should consult the updated extension templates.
new object type: "catalogue", with base class BaseCatalogue. Catalogues contain collections of estimators, metrics, datasets, and/or any other type of object. They can be natively passed to the sktime benchmarking framework to set up pre-defined benchmarks with ease, e.g., estimators from the M4 forecasting competition.
passing an integer fh to forecasters is now interpreted as range(1, fh+1). For example, passing fh=3 will forecast the next 3 periods, and is equivalent to passing fh=[1, 2, 3], or fh=range(1, 4).
forecasters now possess a programmatic API point for pre-training, the pretrain method. Global forecasters, indicated by the capability:pretrain tag, can be pre-trained via pretrain on global training data on which later no forecast is queried via predict. All forecasters have pretrain, but it is a no-op for those with capability:pretrain tag value False.
a new tag capability:non_contiguous_X has been introduced. This boolean tag describes whether a forecaster can handle non-contiguous future-known exogenous data X in predict.
The sktime.libs.pykalman module has been removed in favour of the pykalman package, also maintained by GC.OS and the sktime team. Users of pykalman based estimators in sktime can continue to use them, and the estimator API itself will not change. However, users will need to install pykalman separately, as it is no longer included in sktime.
The ignores-exogeneous-X tag for forecasters is replaced by capability:exogenous. The logic of the tag is flipped, i.e., if capability:exogenous=True, the estimator can handle exogenous variables. Users and maintainers of third party estimators should update their code and estimators to use the new tag.
the univariate-only tag present in multiple estimator types is renamed to capability:multivariate, with boolean flip. If capability:multivariate=True, the estimator can handle multivariate data. Users and maintainers of third party estimators should update their code and estimators to use the new tag.
forecaster and transformation capability tags for exogenous variables and multivariate data, capability:exogenous and capability:multivariate, replace the ignores-exogeneous-X and univariate-only tags, respectively. Values False and True map to True and False, respectively. The old tags are still aliased until 1.1.0, and a warning will be raised when accessing the old tags. From 1.1.0 onwards, an exception will be raised.
Deprecated tags python_dependencies_alias and univariate-metric have now been removed.
the scitype:y tag present in forecasters is renamed to capability:multivariate. If capability:multivariate=True, the estimator can handle multivariate data, if False, it cannot. Values of the old tag map as follows onto the new tag: "univariate" maps onto False, "multivariate" and "both" map onto True. Until 1.1.0, the old tag, scitype:y, can still be used, but will emit a warning. From 1.1.0 onwards, an exception will be raised. Users and maintainers of third party estimators should update their code and estimators to use the new tag.
the deprecated tag capability:pred_var has been removed.
the old estimator type series-annotator for specific detectors has been removed after deprecation period, and replaced by the umbrella tag detector.
all sktime estimators from the skchange package have been merged into sktime.detection proper, after cessation of maintenance by skchange as part of a separate package.
detection and transformations module folder structure has been flattened, most estimators are now directly in a submodule of sktime.detection or sktime.transformations. In particular, the modules sktime.transformations.series and sktime.transformations.panel no longer exist, instead users should import sktime.transformations.series.[module] directly from sktime.transformations.[module]. The old import paths will still continue to work for until the last 1.X version.
[ENH] catalogues in sktime (8749) jgyasu
[ENH] make _safe_import robust against package import errors (9193) fkiraly
[ENH] craft deserialization utility can now also index scikit-learn estimators (9228) fkiraly
[ENH] Removed series-as-features module (9519) RecreationalMath
[ENH] Clean up exogenous capability tag references (9254) junaidaslam2006
[ENH] change dependency error in estimator __init__ to warning (9962) fkiraly
[ENH] move dynamic tag setting to dedicated __dynamic_tags__ method (10140) fkiraly
[ENH] remove old deprecated tag python_dependencies_alias (10298) fkiraly
[ENH] Add _check_estimator_deps to base classes missing soft dependency enforcement (9805) Deep-Axe
[ENH] MeanSquaredLogarithmicError (MSLE) metric (9140) alphaleporus
[ENH] add extension templates for forecasting metrics (9844) NestroyMusoke
[ENH] Allow list, and dict of estimators in benchmark.add (9916) jgyasu
[ENH] Overall Weighted Average (OWA) Score (9111) jgyasu
[ENH] moved PresplitFilesCV and SingleSplit to split module (9519) RecreationalMath
[ENH] Fix constructor dependency handling in catalogues (10295) fkiraly
[ENH] remove old deprecated tag univariate-metric (10298) fkiraly
[ENH] M4 Forecasting Competition Catalogue (9246) jgyasu
[ENH] TSC Bake Off 2017 Catalogue (9138) jgyasu
[ENH] Prevent dataset downloads in catalogue tests (10324) jgyasu
[ENH] Refactor BaseCatalogue (10320) jgyasu
[ENH] Fix capability:exogenous tags for FM forecasters (#8523) (9163) jdhruv555
[ENH] capability:non_contiguous_X tag and skip non-contiguous X tests for affected estimators (9091) EmanAbdelhaleem
[ENH] reduce forecaster coupling to self._y attribute - part 1 (9195) fkiraly
[ENH] ensure test data in test_update_with_exogenous is nonsingular (9204) fkiraly
[ENH] support for neuralforecast new versions (8750) thisisrick25
[ENH] Enable prediction intervals for AutoTS (9203) TenFinges
[ENH] remove "only multivariate" forecaster property (8343) fkiraly
[ENH] add missing capability:non_contiguous_X tags (9213) fkiraly
[ENH] Enable exogenous support for AutoTS (9210) TenFinges
[ENH] Make ForecastingOptunaSearchCV tests compatible without CmaEsSampler (9269) junaidaslam2006
[ENH] TransformedTargetForecaster.predict_proba, via transformed distribution (#9287) (9346) xenonnn4w
[ENH] pretraining and global forecasting API for forecasters (9056) SimonBlanke
[ENH] xLSTM forecaster (8575) vedantag17
[ENH] add pretraining feature for pytorch-based forecasters (9541) SimonBlanke
[ENH] deprecate forecaster scitype:y tag in favour of capability:multivariate (9212) fkiraly
[ENH] skip StatsForecastMSTL testing on negative data with multiplicative seasonality (9807) fkiraly
[ENH] in forecasting tutorial, make pandas calls upwards compatible (9820) fkiraly
[ENH] Add second test param sets for Prophet forecaster (8855) piyushbiraje
[ENH] Chronos 2 foundation model forecaster (9823) priyanshuharshbodhi1
[ENH] update DummyGlobalForecaster tags to 1.0 (9868) fkiraly
[ENH] Lagllama forecaster (9194) pranavvp16
[ENH] fix optuna sampler fixtures in tests (9771) fkiraly
[ENH] Move StackingForecaster test skip from _config to estimator tags (9975) paramsureliya
[ENH] Move SquaringResiduals test skip rule to estimator tag (9866) Kevin23-design
[ENH] simplify IgnoreX __init__ logic (10094) fkiraly
[ENH] remove deprecated cmaes sampler from optuna facing tests (9216) fkiraly
[ENH] TimerForecaster for zero-shot time series forecasting (9744) PewterZz
[ENH] use BaseForecaster direct inheritance for foundation model forecasters (9276) SimonBlanke
[ENH] move more forecaster code to __post_init__ and __dynamic_tags__. (10146) fkiraly
[ENH] reduce forecaster coupling to self._y attribute - part 2 (9202) fkiraly
[ENH] Allow FMs to handle unequal length hierarchical data (10125) jgyasu
[ENH] Timer: vendor architecture code into sktime.libs and fix deps (10327) geetu040
[ENH] declare statsmodels as soft dep of HCrystalBallAdapter (10104) kpal002
[ENH] add tests:vm tag to HCrystalBallAdapter (10243) fkiraly
[ENH] Unequal length TS test and tag for forecasters (10078) jgyasu
[ENH] pandas 3 compatibility fixes (9821) fkiraly
[ENH] mantis forecaster (10266) vedantag17
[ENH] Add params count in _fit for _NeuralForecastAdapter (8428) Faakhir30
[ENH] Interface TSPulse for classification (10278) Faakhir30
[ENH] remove old deprecated tags univariate-metric (10298) fkiraly
[ENH] neuralprophet forecaster (8348) vedantag17
[ENH] uniformize forecasting tuner parameters and docstrings (10247) fkiraly
[ENH] TimesFM 2.x forecasting foundation models (10291) geetu040
[ENH] Bytedance Timer-S1 forecasting foundation model (10315) geetu040
[ENH] temporarily skip failing test for BoxCoxBiasAdjustedForecaster (10338) fkiraly
[ENH] remove _forecaster suffixes in forecasting submodules (10337) fkiraly
[ENH] Update HFTransformersForecaster tags and docstring (10355) fkiraly
[ENH] fix deprecated import in test_hf_transformers_forecaster.py (10369) fkiraly
[ENH] AutoResearchForecaster - agentic improvement of forecaster blueprint (9855) benHeid
[ENH] FalconTST forecasting foundation model (10361) geetu040
[ENH] Kronos forecasting foundation model (10347) geetu040
[ENH] deduplicate ARIMA test skips (10372) fkiraly
[ENH] Flowstate forecasting foundation model by IBM (10329) Faakhir30
[ENH] PatchTSMixer Forecaster by IBM (10330) Faakhir30
[ENH] passing fh: int to forecast range(1, fh + 1) (10378) fkiraly
[ENH] Johansen cointegration (9040) OldPatrick
[ENH] Impulse Response Function/Parameters for Uni/Multivariate Time-Series (9217) OldPatrick
[ENH] in AlignerDTW and AlignerDTWfromDist, add check for incorrect dtw package installation (9214) DCchoudhury15
[ENH] momentfm time series anomaly detection foundation model (9500) Faakhir30
[ENH] remove redundant code in BaseDetector.__init__ (9963) fkiraly
[ENH] Offline univariate Hampel filter detector (9888) SpitFire19
[ENH] remove deprecated scipy Kulsinski distance from tests (10191, 10252) direkkakkar319-ops
[ENH] Natively merge skchange core detection algorithms (#9636) (9719) yash-sangwan
[ENH] add change conditional test skips to new tests in detection module (10359) fkiraly
[ENH] tspulse for anomaly detection (10308) Faakhir30
[ENH] Hidden layer droput uniformization in the deep learning models (9109) abhimanyudalal1
[ENH] Expose dropout parameters in CNTCNetwork and TapNetNetwork (9208) xenonnn4w
[ENH] torch based MCDCNN regressor and classifier (9232) Faakhir30
[ENH] torch based MLP classifier (9042) RecreationalMath
[ENH] torch based LSTMFCN regressor and classifier (9298) Faakhir30
[ENH] torch based InceptionTime regressor and classifier#9305 (9314) Faakhir30
[ENH] Implement torch based tapnet time series classifier and regressor (9412) srupat
[ENH] Make conv layers dynamic in MCDCNNClassifierTorch and MCDCNNRegressorTorch (9342) Faakhir30
[ENH] Make MLP TensorFlow layers dynamic via n_layers and hidden_dim params (9607) varun-kht
[ENH] Make FCN TensorFlow layers dynamic via n_layers, filter_sizes, and kernel_sizes (9618) varun-kht
[ENH] Implement ConvTran deep learning estimator for classification and regression in PyTorch (9550) srupat
[ENH] replace 1 bare except clauses with except Exception (9502) haosenwang1018
[ENH] torch based InceptionTime regressor and classifier#9305 (9314) Faakhir30
[ENH] Remove X.shape checking, single labeled data warnings from TSC and TSR end estimators. (9450) Faakhir30
[ENH] Refactor instantiation and validation of activations in torch models (9622) Faakhir30
[ENH] in torch based TSC and TSR, add support for evaluation metrics from torchmetrics (9390) Faakhir30
[ENH] Move SimpleRNNClassifier test skip config to tags (9559) Spandan-Mishra
[ENH] Move MCDCNNClassifier test skip config to estimator tags (9813) archittmittal
[ENH] torch based MACNN regressor and classifier (9305) Faakhir30
[ENH] torch based CNN regressor and classifier (9330) Faakhir30
[ENH] mantis classifier foundational model (10282) vedantag17
[ENH] TSC Bake Off 2017 Catalogue (9138) jgyasu
[ENH] in torch based TSC and TSR, add support for evaluation metrics from torchmetrics (9390) Faakhir30
[ENH] Complete stale 0.41.0 deprecations for CNN/RNN activation defaults (10207) onkar717
[ENH] add second parameter set for TimeSeriesKernelKMeans (9960) fkiraly
[ENH] torch based RNN regressor (9013) RecreationalMath
[ENH] Hidden layer droput uniformization in the deep learning models (9109) abhimanyudalal1
[ENH] Expose dropout parameters in CNTCNetwork and TapNetNetwork (9208) xenonnn4w
[ENH] torch based MCDCNN regressor and classifier (9232) Faakhir30
[ENH] torch based LSTMFCN regressor and classifier (9298) Faakhir30
[ENH] torch based InceptionTime regressor and classifier#9305 (9314) Faakhir30
[ENH] torch based MLP regressor (9223) Faakhir30
[ENH] Make conv layers dynamic in MCDCNNClassifierTorch and MCDCNNRegressorTorch (9342) Faakhir30
[ENH] Make MLP TensorFlow layers dynamic via n_layers and hidden_dim params (9607) varun-kht
[ENH] Make FCN TensorFlow layers dynamic via n_layers, filter_sizes, and kernel_sizes (9618) varun-kht
[ENH] Implement ConvTran deep learning estimator for classification and regression in PyTorch (9550) srupat
[ENH] replace 1 bare except clauses with except Exception (9502) haosenwang1018
[ENH] torch based InceptionTime regressor and classifier#9305 (9314) Faakhir30
[ENH] Remove X.shape checking, single labeled data warnings from TSC and TSR end estimators. (9450) Faakhir30
[ENH] Refactor instantiation and validation of activations in torch models (9622) Faakhir30
[ENH] in torch based TSC and TSR, add support for evaluation metrics from torchmetrics (9390) Faakhir30
[ENH] Implement torch based tapnet time series classifier and regressor (9412) srupat
[ENH] torch based MACNN regressor and classifier (9305) Faakhir30
[ENH] torch based CNN regressor and classifier (9330) Faakhir30
[ENH] in torch based TSC and TSR, add support for evaluation metrics from torchmetrics (9390) Faakhir30
[ENH] Complete stale 0.41.0 deprecations for CNN/RNN activation defaults (10207) onkar717
[ENH] Allow ColumnwiseTransformer to work directly with sklearn transformers using TabularToSeriesAdaptor (9132) crocmons
[ENH] tsfel feature extraction transformer (9082) Faakhir30
[ENH] tsfeatures feature extraction transformer (9107) Faakhir30
[ENH] DegreeDayFeatures transformer for heating and cooling degree days (9231) zstinnett3
[ENH] Move RandomIntervalSegmenter test skip config to tags (9313) Mohit25f101
[ENH] In WindowSummarizer, replaced deprecated pandas grouper with keys (9367)marrov, adan-shahid
[ENH] Remove deprecated ColumnTransformer (9979) fkiraly
[ENH] remove stale RowTransformer entries from test config (10129) Giggitycountless
[ENH] remove upper numpy bound in MatrixProfileTransformer and address segmentation faults (10339) CloseChoice
[ENH] Flatten transformations module (10371, 10168) fkiraly, CloseChoice
[ENH] estimator tag for extra testing dependencies (9207) fkiraly
[ENH] Prevent shell injection in sktime/utils/git_diff.py (9616) Ironankit525
[ENH] update test_run_test_for_class to take into account VM-only tests (9836) fkiraly
[ENH] update test_run_test_for_class to take into account VM-only tests - part 2 (9865) fkiraly
[ENH] Add tests:vm to pytorch estimators base layer (9542) Faakhir30
[ENH] simplify show_versions by removing alias handling (10356) fkiraly
[ENH] make show_versions normalization proof, e.g., for huggingface-hub (10364) fkiraly
[ENH] complete deprecation of fdiff(mode="full") - replace DeprecationWarning with ValueError (9662) PewterZz
[DOC] Fix broken pytorch_forecasting documentation references (8923) G26karthik
[DOC] callbacks missing in the docstring of some deep learning models (9160) Aniketsy
[DOC] add missing callbacks to FCNClassifier docstring (9183) 12jaspreetsingh
[DOC] correct typo 'occurence' to 'occurrence' (9339) thecaptain789
[DOC] correct typo 'neccessary' to 'necessary' (9349) thecaptain789
[DOC] Typo fixes and docstring cleanup in clustering and distances (9670) Gautam-Bharadwaj
[DOC] fix typo: remove duplicate 'to' in docstring (9683) AyushI7G
[DOC] Programmatic list of estimator types in the documentation (9647) purvanshjoshi
[DOC] Added an explanatory note before deprecation examples (9061) RecreationalMath
[DOC] Fixed Docs and broken hyperlinks (9445) gun29may
[DOC] improve programmatic list of estimator types in API reference (9774) neharoy3
[DOC] Fix np.darray typo to np.ndarray in docs and comments (9826) Saloni-0465
[DOC] Fix typos in docstrings and comments (9673) Ask-812
[DOC] Fix typo in forecasting notebook (choic → choice) (9987) R2-STAR
[DOC] improve formatting in BaseForecaster docstrings (10047) fkiraly
[DOC] shorten KalmanFilterTransformerSIMD to remove pykalman dependency (10057) fkiraly
[DOC] Fix documented default value of error_score parameter in forecasting benchmark (10062) benHeid
[DOC] Fix "to to" double-word typo in docstrings (10077) direkkakkar319-ops
[DOC] improve DirectReductionForecaster docstring (10091) fkiraly
[DOC] improved docstring for craft, deps, and imports serialization utilities (9261) fkiraly
[DOC] Skip fit and predict in doc example of Prophetverse and HierarchicalProphet (10172) jgyasu
[DOC] clarify _post_init__ contract w.r.t. memory and compute heavy operations (10188) fkiraly
[DOC] Fix incorrect reference in SlidingWindowSplitter docstring (9825) Saloni-0465
[DOC] Document __dynamic_tags__ constructor method in extension templates (10296) fkiraly
[DOC] ForecastingRandomizedSearchCV usage example in docstring (10294) SajeelHussain
[DOC] ForecastingGridSearchCV Examples indentation fix in docstring (10303) SajeelHussain
[DOC] add mentions of python 3.14 in missing documentation locations (10317) fkiraly
[DOC] in developer contribution guide, update documentation on pull request stacking (10304) fkiraly
[DOC] Fix WDTW docstring typo (10331) Saloni-0465
[DOC] clarify TemporalTrainTestSplitter docstring (10341) adrynalean
[DOC] Make spelling consistent: "exogeneous" to "exogenous" (10135) direkkakkar319-ops
[MNT] add vm estimators to test-all workflow (9112) neha222222
Bump actions/download-artifact from 6 to 7 (9167) dependabot[bot]
Bump actions/upload-artifact from 5 to 6 (9168) dependabot[bot]
[MNT] bound torch<2.6 (9171) fkiraly
[MNT] CI step with pinned dependencies in August 2024 (9170) fkiraly
[MNT] scipy 1.17 compatibility patch (9238) fkiraly
[ENH] dockerfiles: Python 3.13 and 3.14 support (9130) crocmons
[MNT] [Dependabot](deps-dev): Update sphinx-gallery requirement from <0.20.0 to <0.21.0 (9148) dependabot[bot]
[MNT] bound lightning<2.6 (9173) fkiraly
Update skpro requirement from <2.11.0,>=2 to >=2,<2.12.0 (9179) dependabot[bot]
[MNT] merge release wheels testing workflows (9190) fkiraly
[MNT] prioritize empty return in run_test_vm (9192) fkiraly
[ENH] Prevent test_all to run on forks (9123) Abelarm
[MNT] move autots testing to dedicated VM (9206) fkiraly
Update autots requirement from <0.7,>=0.6.1 to >=0.6.1,<1.1 (9147) dependabot[bot]
[MNT] [Dependabot](deps-dev): Update networkx requirement from <3.5 to <3.7 (9221) dependabot[bot]
[MNT] remove pandas 1 CI (9387) fkiraly
Bump actions/download-artifact from 7 to 8 (9513) dependabot[bot]
Bump actions/upload-artifact from 6 to 7 (9514) dependabot[bot]
[MNT] Updated lower bound for skbase version support (9723) RecreationalMath
[MNT] all-contributors update (9803) github-actions[bot]
[MNT] increase numpy bound to <2.5 (9817) fkiraly
[MNT] clean up depsets for 1.0 release (9386) fkiraly
[MNT] simplify all_extras soft dependency set (9870) fkiraly
[MNT] Use uv date cutoffs instead of extras for old dependency version testing (9872)zanieb, fkiraly
Update skpro requirement from <2.12.0,>=2 to >=2,<2.13.0 (9886) dependabot[bot]
[MNT] [Dependabot](deps-dev): Update sphinx-issues requirement from <6.0.0 to <7.0.0 (9907) dependabot[bot]
Update scikit-base requirement from <0.14.0,>=0.13.1 to >=0.13.1,<1.1.0 (9952) dependabot[bot]
Update pytest-randomly requirement from <4.1,>=3.15 to >=3.15,<4.2 (9961) dependabot[bot]
[MNT] make test-est CI element require test-nosoftdeps (9974) fkiraly
[MNT] all-contributors update (9927) github-actions[bot]
[MNT] Update CODEOWNERS to active core developers (9978) fkiraly
[MNT] point pykalman import to package after its restoration (9976) fkiraly
[MNT] remove module specific tests from CI (10011) fkiraly
[MNT] chunk estimator VM tests to ensure maximum matrix size of 256 is not exceeded in test-est (10012) fkiraly
[MNT] [Dependabot](deps-dev): Update sphinx-gallery requirement from <0.21.0 to <0.22.0 (10032) dependabot[bot]
[MNT] in test-est CI, ensure all dependencies are resolved at once (10046) fkiraly
[MNT] all-contributors update (10033) github-actions[bot]
[MNT] all-contributors update (10149) github-actions[bot]
[MNT] add the artificial-intelligence keyword to pyproject.toml (10242) fkiraly
[MNT] Update Makefile install target for pyproject builds (10244) nvphungdev
[MNT] replace deprecated setup.py install with pip install in Makefile (10254) algojogacor
[MNT] all-contributors update (10334) github-actions[bot]
Update skpro requirement from <2.13.0,>=2 to >=2,<2.14.0 (10353) dependabot[bot]
[MNT] remove _contrib folder from package (10370) fkiraly
[MNT] all-contributors update (9182) github-actions[bot]
[MNT] all-contributors update (10210) github-actions[bot]
[MNT] Deduplicate sktime.utils.dependencies methods (9403) RecreationalMath
[MNT] Update Makefile install target for pyproject builds (10244) nvphungdev
[MNT] 1.0.0 deprecations and change actions (10377) fkiraly
[BUG] fixes to craft utility and datatypes returns (9688) fkiraly
[BUG] Add missing raise keywords to Exceptions across core modules (10060) rakshaak29
[BUG] fix cutoff in evaluate global mode, use y_hist not global y_train (10373) Nischal1425
[BUG] Fix load_fpp3 by updating CRAN package version strings (10147) Rusheel86
[BUG] fixes to craft utility and datatypes returns (9688) fkiraly
[BUG] fix gluonts_ListDataset_panel metadata (10061) goyaladitya05
[BUG] Fix hf_xet dependency name to hf-xet in moirai_forecaster.py (9189) gupta-tilak
[BUG] fix VECM wrong tag and internal variable name handling (9205) fkiraly
[BUG] Skip non-converged AutoETS fits during model selection (9292) sabasiddique1
[BUG] bound transformers dependency (9596) SimonBlanke
[BUG] Resolve Prophet crash by robustly handling stan_backend and fixing test dimensions (9234) DCchoudhury15
[BUG] fix API non-conformance in Chronos2Forecaster (9924) fkiraly
[BUG] Chronos2 truncation axis fix and added a regression test (9955) Chetansahney
[BUG] fix LagLlamaForecaster._predict self-mutation of _is_range_index (9921) shaun0927
[BUG] Fix PatchTSTForecaster compatibility with transformers>=5.0 (9874) neha222222
[BUG] Fix TinyTimeMixerForecaster AttributeError with transformers>=5.0 (9875) neha222222
[BUG] fix GreykiteForecaster test flags (10150) fkiraly
[BUG] ChronosForecaster compatibility for transformers 5.X (10132) nikol-damyanova
[BUG] Fix Chronos-Bolt incompatibility with transformers >=5.x and restore test coverage (10167) Sonika-19
[BUG] Fix error message for forecaster count validation for TransformedTargetForecaster (10043) benHeid
[BUG] Fix TimesFM dependency imports and enable VM testing (10265) geetu040
[BUG] fix TimesFMForecaster multiton method name (10302) fkiraly
[BUG] fix pmdarima forecasters y.name preservation in predict (10276) fkiraly
[BUG] add missing tests:libs and tests:vm tags for forecasters with internal dependencies (10277) fkiraly
[BUG] fix HFTransformersForecaster compatibility with transformers >= 4.41 (10087) Nischal1425
[BUG] fix predict_interval index mismatch for ARIMA / AutoARIMA full in-sample fh (10319) kpal002
[BUG] Fix ChronosForecaster T5Stack initialization on transformers>=5.0 (10085) sabasiddique1
[ENH] fork necessary functions from seasonal package (9155) JATAYU000
[BUG] Fix IndexError in twe_alignment_path default bounding matrix (9196) DCchoudhury15
[BUG] Remove duplicate BasetDetector._transform_score function (9835) direkkakkar319-ops
[BUG] fix BaseDetector sparse-to-dense conversion for pd.DataFrame input (9890) direkkakkar319-ops
[BUG] Fix NaN handling in ProximityForest _stdp calculation (9186) TenFinges
[BUG] Improve default hidden layer activation in CNN based classifier and regressor (9050) RecreationalMath
[BUG] Fix padding macnn conv layer tensorflow (9307) Faakhir30
[BUG] fix optimizer instantiation in BaseDeepClassifierPyTorch class (9359) RecreationalMath
[BUG] RocketClassifier - fix expected outputs + readd test (fixes issue #7921) (9152) SimonBlanke
[BUG] Followup bug fixing on Inceptiontime (9643) Faakhir30
[BUG] Fix TSCOptCV incorrect base class inheritance (10098) direkkakkar319-ops
[BUG] Fix parameter order in MCDCNN estimators' docstring and constructor & add type hints (9393) Faakhir30
[BUG] Improve default hidden layer activation in CNN based classifier and regressor (9050) RecreationalMath
[BUG] Fix padding macnn conv layer tensorflow (9307) Faakhir30
[BUG] Followup bug fixing on Inceptiontime (9643) Faakhir30
[BUG] Fix parameter order in MCDCNN estimators' docstring and constructor & add type hints (9393) Faakhir30
[BUG] Fix ColumnwiseTransformer sklearn support using coerce_scitype (9200) DCchoudhury15
[BUG] Fix WindowSummarizer bfill across multiindex groups (9243) marrov
[BUG] Guard logsignature post-rescaling and add regression test (9285) sabasiddique1
[BUG] Replace .rename() with .rename_categories() to prevent FutureWarning in DateTimeFeatures (9433) FlyingDragon112
[BUG] Fix TSBootstrap test tag typo tests:vms -> tests:vm (9959) PragnyaKhandelwal
[BUG] remove eval from TransformIf condition evaluation (9993) PragnyaKhandelwal
[BUG] Fix duplicated mean in std_transformer (10186) Saloni-0465
[BUG] fix mlflow test failure - double use of database reference (9185) jdhruv555
[BUG] Failing test case in fracdiff (10185) direkkakkar319-ops
12jaspreetsingh, Abelarm, abhimanyudalal1, adan-shahid, adrynalean, algojogacor, alphaleporus, Aniketsy, archittmittal, Ask-812, AyushI7G, benHeid, Chetansahney, CloseChoice, crocmons, DCchoudhury15, Deep-Axe, direkkakkar319-ops, EmanAbdelhaleem, Faakhir30, fkiraly, FlyingDragon112, G26karthik, Gautam-Bharadwaj, geetu040, Giggitycountless, goyaladitya05, gun29may, gupta-tilak, haosenwang1018, Ironankit525, JATAYU000, jdhruv555, jgyasu, junaidaslam2006, Kevin23-design, kpal002, marrov, Mohit25f101, neha222222, neharoy3, NestroyMusoke, nikol-damyanova, Nischal1425, nvphungdev, OldPatrick, onkar717, paramsureliya, PewterZz, piyushbiraje, PragnyaKhandelwal, pranavvp16, priyanshuharshbodhi1, purvanshjoshi, R2-STAR, rakshaak29, RecreationalMath, Rusheel86, sabasiddique1, SajeelHussain, Saloni-0465, shaun0927, SimonBlanke, Sonika-19, Spandan-Mishra, SpitFire19, srupat, TenFinges, thecaptain789, thisisrick25, varun-kht, vedantag17, xenonnn4w, yash-sangwan, zstinnett3
Mini-release to correct Zenodo record.
Mini-release to correct Zenodo record.
Please see our changelog for a description of all changes.
Full Changelog: https://github.com/sktime/sktime/compare/v0.40.0...v0.40.1
Feature release and python 3.14 compatibility.
Feature release and python 3.14 compatibility.
Please see our changelog for a description of all changes.
@Abelarm, @AdKnow, @Akai01, @alphaleporus, @AryanDhanuka10, @Astrael1, @Faakhir30, @fkiraly, @JATAYU000, @jgyasu, @leonicus, @mabuimo, @martinloretzzz, @Navya-1817, @noxthot, @omkar-334, @poopsiclepooding, @pranavvp16, @Priyanshu1303d, @RecreationalMath, @Sanchay117, @sanskarmodi8, @SimonBlanke, @swetha3456, @Tanuj-Taneja1, @XAheli, @yarnabrina
Full Changelog: https://github.com/sktime/sktime/compare/v0.39.0...v0.40.0
sktime now supports python 3.14 ( #8939 ) @fkiraly
ARAR forecaster ( #9062 ) @Akai01
Double ML forecaster for causal forecasting with confounder adjustment ( #8797 ) @geetu040 , @XAheli
Activation functions are now configurable in most deep learning regressors and classifiers ( #8934 ) @noxthot
ConvTimeNet for time series forecasting ( #8963 ) @Tanuj-Taneja1
TSB forecaster for intermittent demand ( #7995 ) @swetha3456
TiRexForecaster now uses the official tirex-ts package ( #8876 ) @martinloretzzz
fABBA Transform ( #8838 ) @poopsiclepooding
pmdarima based estimators now support numpy 2.X , since pmdarima is now compatible with it.
python 3.14 is now supported by the sktime framework. Individual estimators may not support it yet, users should consult the python_dependencies tag.
pytorch-forecasting (forecasting soft dependency) bounds have been updated to >=1.0.0,<1.6.0 .
freia and pykan soft dependencies have been removed from the deep learning soft dependency set dl . Users of estimators depending on these packages will need to install them separately.
[ENH] extend scitype type inference utility to cover sklearn metrics and splitters ( #8883 ) @fkiraly
[ENH] Test param sets for MeanAbsoluteError ( #8957 ) @Astrael1
[ENH] test parameters for MAE ( #8924 ) @leonicus
[ENH] refactor benchmark storage handlers ( #8940 ) @omkar-334
[ENH] benchmarks to work without file being written ( #8937 ) @fkiraly
[ENH] internal architecture improvements to benchmarking module ( #8250 ) @fkiraly
[ENH] in benchmarking module, replace “no file IO” handling with null handler ( #8977 ) @fkiraly
[ENH] MeanArctangentAbsolutePercentageError (MAAPE) metric (#9094) ( #9095 ) @alphaleporus
[ENH] only run test_load_fpp3_public when rdata is installed ( #8891 ) @noxthot
[ENH] More Dataset Classes for Existing Loaders ( #8894 ) @jgyasu
[ENH] TSB forecaster for intermittent demand ( #7995 ) @swetha3456
[ENH] Moved SARIMAX skipping all tests from _config.py to estimator tag ( #8981 ) @mabuimo
[ENH] ConvTimeNet for time series forecasting ( #8963 ) @Tanuj-Taneja1
[ENH] Use tirex-ts package for TiRex model ( #8876 ) @martinloretzzz
[ENH] improve soft dependency isolation patterns ( #8997 ) @fkiraly
[ENH] DMLForecaster for causal forecasting with confounder adjustment ( #8797 ):user: geetu040 , @XAheli
[ENH] update MOIRAIForecaster with hf_xet requirement ( #9032 ) @fkiraly
[ENH] Align _predict method parameters {fh, X} order across BaseForecaster subclasses ( #9066 ) @Faakhir30
[ENH] ARAR forecaster ( #9062 ) @Akai01
[ENH] UnobservedComponents second set of test parameters ( #8970 ) @Abelarm
[ENH] detector test scenario and smoke test for detector estimators #8896 ( #8914 ) @AryanDhanuka10
[ENH] Test param sets for InformationGainSegmentation ( #8968 ) @Astrael1
[ENH] set some missing estimator tags ( #9072 ) @fkiraly
[ENH] BOSSEnsemble second set of test parameters ( #8922 ) @Abelarm
[ENH] MrSQM - correctly set randomness capability tags ( #8941 ) @fkiraly
[ENH] Moved Proximity classes skipped test from tests._config to estimator tags ( #8935 ) @Abelarm
[ENH] CanonicalIntervalForest second set of test parameters ( #8951 ) @Abelarm
[ENH] Propagate and unify use of activation and activation_* in various deep learning based models ( #8934 ) @noxthot
[ENH] torch based RNN classifier ( #8842 ) @RecreationalMath
[ENH] correctly set random seed tags for MVTSTransformerClassifier ( #9016 ) @fkiraly
[ENH] improve soft dependency isolation patterns ( #8997 ) @fkiraly
[ENH] Propagate and unify use of activation and activation_* in various deep learning based models ( #8934 ) @noxthot
[ENH] Move SimpleRNNRegressor test skip config from tests._config to estimator tags ( #8982 ) @AdKnow
[ENH] Move test skip config from tests._config to FCN Regressor tags ( #9084 ) @Navya-1817
[ENH] Extends HolidayFeature for PeriodIndex ( #8777 ) @JATAYU000
[ENH] Test parameter sets for ReducerTransform ( #8911 ) @Priyanshu1303d
[ENH] Test parameter sets for TimeBinner ( #8865 ) @AryanDhanuka10
[ENH] Test parameter sets for LogTransformer ( #8956 ) @Astrael1
[ENH] transformer tags for changes to unequal_length property ( #8959 ) @fkiraly
[ENH] fABBA Transform ( #8838 ) @poopsiclepooding
[ENH] Test parameter sets for TruncationTransformer ( #8948 ) @Astrael1
[ENH] Make TruncationTransformer handle incorrect arguments ( #8980 ) @Astrael1
[ENH] Test parameter sets for DWTTransformer ( #8983 ) @mabuimo
[ENH] Test parameter sets for WhiteNoiseAugmenter ( #8943 ) @Astrael1
[ENH] set some missing estimator tags ( #9072 ) @fkiraly
[ENH] Move RandomIntervals test skip from _config.py to estimator tags ( #8874 ) @sanskarmodi8
[ENH] remove SeasonalityPeriodogram from framework core estimator tests ( #8893 ) @fkiraly
[ENH] skip ClearSky categorical input test ( #8926 ) @fkiraly
[ENH] detector test scenario and smoke test for detector estimators ( #8914 ) @AryanDhanuka10
[ENH] Move PlateauFinder test skip from config to estimator tags ( #8919 ) @Navya-1817
[ENH] add helpful error message to check_estimator if estimator dependencies are not present ( #8958 ) @fkiraly
[ENH] is_module_changed to return True if path cannot be found instead of exception raised ( #8991 ) @fkiraly
[DOC] Fix class name of UCR UEA datasets for docs build ( #8859 ) @jgyasu
[DOC] fix formatting of DateTimeFeatures docstring ( #8927 ) @fkiraly
[DOC] Add mention of lowerdeps in testing guide ( #8930 ) @Astrael1
[DOC] Fix code / doc mismatches in regards to activation for TapNetRegressor , CNTCNetwork , and MACNNRegressor ( #8899 ) @noxthot
[DOC] correct extension templates re capability:exogenous ( #8920 ) @fkiraly
[DOC] minor formatting improvement to ForecastingHorizon docstring ( #8960 ) @fkiraly
[DOC] improved developer documentation for adding new base interfaces ( #9003 ) @fkiraly
[DOC] documented tags for time series aligners ( #9000 ) @fkiraly
[DOC] better documentation for benchmarking file backend handlings json, parquet, csv. ( #9043 ) @fkiraly
[DOC] Adding some links in example notebook ( #8925 ) @Astrael1
[DOC] Add Classifier Benchmarking Notebook to Docs ( #8901 ) @jgyasu
[DOC] Fix rendering of examples of AlignerDTW ( #9067 ) @Faakhir30
[DOC] Add seasonal example to NaiveForecaster (#9085) ( #9086 ) @alphaleporus
[MNT] address setuptools vulnerabilities ( #8887 ) @yarnabrina
[MNT] in tests, set matplotlib silent backend and print python deps ( #8908 ) @fkiraly
[MNT] Change allcontributors update to a CRON job ( #8849 ) @JATAYU000
[MNT] Skip doctests where external dataset is downloaded ( #8915 ) @jgyasu
[MNT] Update skforecast bounds ( #8888 ) @yarnabrina
[MNT] all-contributors update ( #8962 ) @github-actions[bot]
[MNT] temporarily skip failing ProximityTree and ProximityForest tests ( #8947 ) @fkiraly
[MNT] Move CI from pip to uv ( #8918 ) @Abelarm
[MNT] Bump astral-sh/setup-uv from 6 to 7 ( #8972 ) @dependabot[bot]
[MNT] Bump actions/setup-node from 5 to 6 ( #8974 ) @dependabot[bot]
[MNT] all-contributors update ( #8979 ) @github-actions[bot]
[MNT] revert to pip on failing test_external_data ( #8985 ) @fkiraly
[MNT] Fixing CI test failing for python 3.10 ( #8986 ) @Abelarm
[MNT] Update pytorch-forecasting requirement from <1.5.0,>=1.0.0 to >=1.0.0,<1.6.0 ( #8961 ) @dependabot[bot]
[MNT] Bump actions/download-artifact from 5 to 6 ( #8996 ) @dependabot[bot]
[MNT] Bump actions/upload-artifact from 4 to 5 ( #8995 ) @dependabot[bot]
[MNT] all-contributors update ( #8998 ) @github-actions[bot]
[MNT] remove pykan soft dependency from dl depset ( #9001 ) @fkiraly
[MNT] remove freia soft dependency from dl depset ( #9002 ) @fkiraly
[MNT] temporary skip of failing tests in ClaspSegmentation ( #9015 ) @fkiraly
[MNT] add new AI pypi classifier to pyproject.toml ( #9024 ) @fkiraly
[MNT] all-contributors update ( #9029 ) @github-actions[bot]
[MNT] bump prophetverse facing placeholder record version bounds to prophetverse<0.11.0 ( #8992 ) @fkiraly
[MNT] isolate peft in HFTransformersForecaster doctest ( #9035 ) @fkiraly
[MNT] Update pytest requirement from <8.5,>=7.4 to >=7.4,<9.1 ( #9054 ) @dependabot[bot]
[MNT] Add funding.json ( #9057 ) @jgyasu
[MNT] fix funding.json format ( #9063 ) @fkiraly
[MNT] Fix funding.json ( #9064 ) @fkiraly
[MNT] Add CITATION.cff for sktime ( #9051 ) @jgyasu
[MNT] python 3.14 support ( #8939 ) @fkiraly
[MNT] Bump actions/checkout from 5 to 6 ( #9090 ) @dependabot[bot]
[MNT] fix set-output deprecation in github-actions ( #8913 ) @noxthot
[MNT] all-contributors update ( #9046 ) @github-actions[bot]
[MNT] all-contributors update ( #9065 ) @github-actions[bot]
[MNT] all-contributors update ( #9096 ) @github-actions[bot]
[MNT] 0.40.0 deprecations and change actions ( #9080 ) @fkiraly
[BUG] Fix Dataset Loader Loading Logic in Benchmarking Module ( #8861 ) @jgyasu
[BUG] Correct task_id in benchmark result for all kinds of dataset loaders ( #8897 ) @jgyasu
[BUG] fix DatasetDownloader doctest ( #8860 ) @fkiraly
[BUG] fix USChange dataset class ( #9052 ) @fkiraly
[BUG] Fix ChronosForecaster failure on newer transformers versions ( #8877 ) @Sanchay117
[BUG] fixed probabilistic predictions of SkforecastRecursive ( #8921 ) @yarnabrina
[BUG] ChronosForecaster gives a “Can’t pickle” error ( #8944 ) @JATAYU000
[BUG] Fix ConvTimeNetForecaster test configuration paths ( #8990 ) @fkiraly
[BUG] Fix _safe_import import error in ConvTimeNetForecaster ( #9012 ) @jgyasu
[BUG] Enable exogenous variables support in MOIRAIForecaster ( #9008 ) @pranavvp16
[BUG] MSTL does not support a trend_forecaster that needs fh during fit ( #8931 ) @Abelarm
[BUG] BaggingForecaster fails test_update_with_exogenous_variables ( #9053 ) @JATAYU000
[BUG] Fix ForecastingPipeline MultiIndex handling in _transform method ( #9075 ) @XAheli
[BUG] ESRNNForecaster fails test_persistence_via_pickle ( #8989 ) @JATAYU000
[BUG] Fix SubLOF mutating input DataFrame in _predict ( #8873 ) @sanskarmodi8
[BUG] fix GreedyGaussianSegmentation predict API conformance ( #8967 ) @Astrael1
[BUG] in CNNClassifier and CNNRegressor , fix hidden layer activation function parameter passing ( #8880 ) @noxthot
[BUG] fix proximity forest random-state + pickle ( #8965 ) @SimonBlanke
[BUG] fix ColumnEnsembleClassifier remainder handling ( #8916 ) @Abelarm
[BUG] Improve default activation_hidden in SimpleRNNClassifier , SimpleRNNRegressor & RNNNetwork ( #9048 ) @RecreationalMath
[BUG] in CNNClassifier and CNNRegressor , fix hidden layer activation function parameter passing ( #8880 ) @noxthot
[BUG] remove KNeighborsTimeSeriesRegressor spurious numba dependency tag ( #9019 ) @fkiraly
[BUG] Add missing estimator dependency checks in BaseRegressor class ( #9018 ) @RecreationalMath
[BUG] Improve default activation_hidden in SimpleRNNClassifier , SimpleRNNRegressor & RNNNetwork ( #9048 ) @RecreationalMath
@Abelarm , @AdKnow , @Akai01 , @alphaleporus , @AryanDhanuka10 , @Astrael1 , @Faakhir30 , @fkiraly , @JATAYU000 , @jgyasu , @leonicus , @mabuimo , @martinloretzzz , @Navya-1817 , @noxthot , @omkar-334 , @poopsiclepooding , @pranavvp16 , @Priyanshu1303d , @RecreationalMath , @Sanchay117 , @sanskarmodi8 , @SimonBlanke , @swetha3456 , @Tanuj-Taneja1 , @XAheli , @yarnabrina
sktime now supports python 3.14 (8939) fkiraly
ARAR forecaster (9062) Akai01
Double ML forecaster for causal forecasting with confounder adjustment (8797) geetu040, XAheli
Activation functions are now configurable in most deep learning regressors and classifiers (8934) noxthot
ConvTimeNet for time series forecasting (8963) Tanuj-Taneja1
TSB forecaster for intermittent demand (7995) swetha3456
TiRexForecaster now uses the official tirex-ts package (8876) martinloretzzz
fABBA Transform (8838) poopsiclepooding
pmdarima based estimators now support numpy 2.X, since pmdarima is now compatible with it.
python 3.14 is now supported by the sktime framework. Individual estimators may not support it yet, users should consult the python_dependencies tag.
pytorch-forecasting (forecasting soft dependency) bounds have been updated to >=1.0.0,<1.6.0.
freia and pykan soft dependencies have been removed from the deep learning soft dependency set dl. Users of estimators depending on these packages will need to install them separately.
[ENH] extend scitype type inference utility to cover sklearn metrics and splitters (8883) fkiraly
[ENH] Test param sets for MeanAbsoluteError (8957) Astrael1
[ENH] test parameters for MAE (8924) leonicus
[ENH] refactor benchmark storage handlers (8940) omkar-334
[ENH] benchmarks to work without file being written (8937) fkiraly
[ENH] internal architecture improvements to benchmarking module (8250) fkiraly
[ENH] in benchmarking module, replace "no file IO" handling with null handler (8977) fkiraly
[ENH] MeanArctangentAbsolutePercentageError (MAAPE) metric (#9094) (9095) alphaleporus
[ENH] only run test_load_fpp3_public when rdata is installed (8891) noxthot
[ENH] More Dataset Classes for Existing Loaders (8894) jgyasu
[ENH] TSB forecaster for intermittent demand (7995) swetha3456
[ENH] Moved SARIMAX skipping all tests from _config.py to estimator tag (8981) mabuimo
[ENH] ConvTimeNet for time series forecasting (8963) Tanuj-Taneja1
[ENH] Use tirex-ts package for TiRex model (8876) martinloretzzz
[ENH] improve soft dependency isolation patterns (8997) fkiraly
[ENH] DMLForecaster for causal forecasting with confounder adjustment (8797)geetu040, XAheli
[ENH] update MOIRAIForecaster with hf_xet requirement (9032) fkiraly
[ENH] Align _predict method parameters {fh, X} order across BaseForecaster subclasses (9066) Faakhir30
[ENH] ARAR forecaster (9062) Akai01
[ENH] UnobservedComponents second set of test parameters (8970) Abelarm
[ENH] detector test scenario and smoke test for detector estimators #8896 (8914) AryanDhanuka10
[ENH] Test param sets for InformationGainSegmentation (8968) Astrael1
[ENH] set some missing estimator tags (9072) fkiraly
[ENH] BOSSEnsemble second set of test parameters (8922) Abelarm
[ENH] MrSQM - correctly set randomness capability tags (8941) fkiraly
[ENH] Moved Proximity classes skipped test from tests._config to estimator tags (8935) Abelarm
[ENH] CanonicalIntervalForest second set of test parameters (8951) Abelarm
[ENH] Propagate and unify use of activation and activation_* in various deep learning based models (8934) noxthot
[ENH] torch based RNN classifier (8842) RecreationalMath
[ENH] correctly set random seed tags for MVTSTransformerClassifier (9016) fkiraly
[ENH] improve soft dependency isolation patterns (8997) fkiraly
[ENH] Propagate and unify use of activation and activation_* in various deep learning based models (8934) noxthot
[ENH] Move SimpleRNNRegressor test skip config from tests._config to estimator tags (8982) AdKnow
[ENH] Move test skip config from tests._config to FCN Regressor tags (9084) Navya-1817
[ENH] Extends HolidayFeature for PeriodIndex (8777) JATAYU000
[ENH] Test parameter sets for ReducerTransform (8911) Priyanshu1303d
[ENH] Test parameter sets for TimeBinner (8865) AryanDhanuka10
[ENH] Test parameter sets for LogTransformer (8956) Astrael1
[ENH] transformer tags for changes to unequal_length property (8959) fkiraly
[ENH] fABBA Transform (8838) poopsiclepooding
[ENH] Test parameter sets for TruncationTransformer (8948) Astrael1
[ENH] Make TruncationTransformer handle incorrect arguments (8980) Astrael1
[ENH] Test parameter sets for DWTTransformer (8983) mabuimo
[ENH] Test parameter sets for WhiteNoiseAugmenter (8943) Astrael1
[ENH] set some missing estimator tags (9072) fkiraly
[ENH] Move RandomIntervals test skip from _config.py to estimator tags (8874) sanskarmodi8
[ENH] remove SeasonalityPeriodogram from framework core estimator tests (8893) fkiraly
[ENH] skip ClearSky categorical input test (8926) fkiraly
[ENH] detector test scenario and smoke test for detector estimators (8914) AryanDhanuka10
[ENH] Move PlateauFinder test skip from config to estimator tags (8919) Navya-1817
[ENH] add helpful error message to check_estimator if estimator dependencies are not present (8958) fkiraly
[ENH] is_module_changed to return True if path cannot be found instead of exception raised (8991) fkiraly
[DOC] Fix class name of UCR UEA datasets for docs build (8859) jgyasu
[DOC] fix formatting of DateTimeFeatures docstring (8927) fkiraly
[DOC] Add mention of lowerdeps in testing guide (8930) Astrael1
[DOC] Fix code / doc mismatches in regards to activation for TapNetRegressor, CNTCNetwork, and MACNNRegressor (8899) noxthot
[DOC] correct extension templates re capability:exogenous (8920) fkiraly
[DOC] minor formatting improvement to ForecastingHorizon docstring (8960) fkiraly
[DOC] improved developer documentation for adding new base interfaces (9003) fkiraly
[DOC] documented tags for time series aligners (9000) fkiraly
[DOC] better documentation for benchmarking file backend handlings json, parquet, csv. (9043) fkiraly
[DOC] Adding some links in example notebook (8925) Astrael1
[DOC] Add Classifier Benchmarking Notebook to Docs (8901) jgyasu
[DOC] Fix rendering of examples of AlignerDTW (9067) Faakhir30
[DOC] Add seasonal example to NaiveForecaster (#9085) (9086) alphaleporus
[MNT] address setuptools vulnerabilities (8887) yarnabrina
[MNT] in tests, set matplotlib silent backend and print python deps (8908) fkiraly
[MNT] Change allcontributors update to a CRON job (8849) JATAYU000
[MNT] Skip doctests where external dataset is downloaded (8915) jgyasu
[MNT] Update skforecast bounds (8888) yarnabrina
[MNT] all-contributors update (8962) github-actions[bot]
[MNT] temporarily skip failing ProximityTree and ProximityForest tests (8947) fkiraly
[MNT] Move CI from pip to uv (8918) Abelarm
[MNT] Bump astral-sh/setup-uv from 6 to 7 (8972) dependabot[bot]
[MNT] Bump actions/setup-node from 5 to 6 (8974) dependabot[bot]
[MNT] all-contributors update (8979) github-actions[bot]
[MNT] revert to pip on failing test_external_data (8985) fkiraly
[MNT] Fixing CI test failing for python 3.10 (8986) Abelarm
[MNT] Update pytorch-forecasting requirement from <1.5.0,>=1.0.0 to >=1.0.0,<1.6.0 (8961) dependabot[bot]
[MNT] Bump actions/download-artifact from 5 to 6 (8996) dependabot[bot]
[MNT] Bump actions/upload-artifact from 4 to 5 (8995) dependabot[bot]
[MNT] all-contributors update (8998) github-actions[bot]
[MNT] remove pykan soft dependency from dl depset (9001) fkiraly
[MNT] remove freia soft dependency from dl depset (9002) fkiraly
[MNT] temporary skip of failing tests in ClaspSegmentation (9015) fkiraly
[MNT] add new AI pypi classifier to pyproject.toml (9024) fkiraly
[MNT] all-contributors update (9029) github-actions[bot]
[MNT] bump prophetverse facing placeholder record version bounds to prophetverse<0.11.0 (8992) fkiraly
[MNT] isolate peft in HFTransformersForecaster doctest (9035) fkiraly
[MNT] Update pytest requirement from <8.5,>=7.4 to >=7.4,<9.1 (9054) dependabot[bot]
[MNT] Add funding.json (9057) jgyasu
[MNT] fix funding.json format (9063) fkiraly
[MNT] Fix funding.json (9064) fkiraly
[MNT] Add CITATION.cff for sktime (9051) jgyasu
[MNT] python 3.14 support (8939) fkiraly
[MNT] Bump actions/checkout from 5 to 6 (9090) dependabot[bot]
[MNT] fix set-output deprecation in github-actions (8913) noxthot
[MNT] all-contributors update (9046) github-actions[bot]
[MNT] all-contributors update (9065) github-actions[bot]
[MNT] all-contributors update (9096) github-actions[bot]
[MNT] 0.40.0 deprecations and change actions (9080) fkiraly
[BUG] Fix Dataset Loader Loading Logic in Benchmarking Module (8861) jgyasu
[BUG] Correct task_id in benchmark result for all kinds of dataset loaders (8897) jgyasu
[BUG] fix DatasetDownloader doctest (8860) fkiraly
[BUG] fix USChange dataset class (9052) fkiraly
[BUG] Fix ChronosForecaster failure on newer transformers versions (8877) Sanchay117
[BUG] fixed probabilistic predictions of SkforecastRecursive (8921) yarnabrina
[BUG] ChronosForecaster gives a "Can't pickle" error (8944) JATAYU000
[BUG] Fix ConvTimeNetForecaster test configuration paths (8990) fkiraly
[BUG] Fix _safe_import import error in ConvTimeNetForecaster (9012) jgyasu
[BUG] Enable exogenous variables support in MOIRAIForecaster (9008) pranavvp16
[BUG] MSTL does not support a trend_forecaster that needs fh during fit (8931) Abelarm
[BUG] BaggingForecaster fails test_update_with_exogenous_variables (9053) JATAYU000
[BUG] Fix ForecastingPipeline MultiIndex handling in _transform method (9075) XAheli
[BUG] ESRNNForecaster fails test_persistence_via_pickle (8989) JATAYU000
[BUG] Fix SubLOF mutating input DataFrame in _predict (8873) sanskarmodi8
[BUG] fix GreedyGaussianSegmentation predict API conformance (8967) Astrael1
[BUG] in CNNClassifier and CNNRegressor, fix hidden layer activation function parameter passing (8880) noxthot
[BUG] fix proximity forest random-state + pickle (8965) SimonBlanke
[BUG] fix ColumnEnsembleClassifier remainder handling (8916) Abelarm
[BUG] Improve default activation_hidden in SimpleRNNClassifier, SimpleRNNRegressor & RNNNetwork (9048) RecreationalMath
[BUG] in CNNClassifier and CNNRegressor, fix hidden layer activation function parameter passing (8880) noxthot
[BUG] remove KNeighborsTimeSeriesRegressor spurious numba dependency tag (9019) fkiraly
[BUG] Add missing estimator dependency checks in BaseRegressor class (9018) RecreationalMath
[BUG] Improve default activation_hidden in SimpleRNNClassifier, SimpleRNNRegressor & RNNNetwork (9048) RecreationalMath
Abelarm, AdKnow, Akai01, alphaleporus, AryanDhanuka10, Astrael1, Faakhir30, fkiraly, JATAYU000, jgyasu, leonicus, mabuimo, martinloretzzz, Navya-1817, noxthot, omkar-334, poopsiclepooding, pranavvp16, Priyanshu1303d, RecreationalMath, Sanchay117, sanskarmodi8, SimonBlanke, swetha3456, Tanuj-Taneja1, XAheli, yarnabrina
changes to tag framework, testing framework
Release focusing on:
Also includes new estimators and feature improvements.
Please see our changelog for a description of all changes.
@Astrael1, @ericjb, @felipeangelimvieira, @fkiraly, @JATAYU000, @jgyasu, @oresthes, @piyushbiraje, @RobKuebler, @SimonBlanke, @sinemkilicdere, @Tanuj-Taneja1, @tingiskhan, @yarnabrina
Full Changelog: https://github.com/sktime/sktime/compare/v0.38.5...v0.39.0
Release focusing on:
python 3.9 end-of-life
changes to tag framework, testing framework
scheduled deprecations
Also includes new estimators relying on changes to tag and testing framework.
TiRex foundation model forecaster ( #8683 ) @sinemkilicdere
Toto foundation model forecaster ( #8708 ) @JATAYU000
Dataset Loader Class to load UCR UEA Datasets from sktime huggingface repo ( #8786 ) @jgyasu
Dataset Loader Class to load Monash Forecasting Datasets from sktime huggingface repo ( #8751 ) @jgyasu
hyperactive based tuners for forecasting and time series classification ( #8705 ) @fkiraly , @SimonBlanke
prophetverse hierarchical forecasters updated; new configurable API, intermittent/hurdle distributions ( #8792 ) @felipeangelimvieira , @tingiskhan
ConvTimeNet for time series classification ( #8761 ) @Tanuj-Taneja1
Element-wise arithmetic operator transformer ( #8711 ) @oresthes
python 3.9 is no longer supported; minimum python version is now 3.10 .
numba bound, where present in dependency sets, has been updated to <0.63 .
the following packages have been removed from dependency sets due to being unmaintained, rarely used, or subject to strict implied secondary dependency bounds: dtaidistance , filterpy , mne , seasonal , stumpy , tbats . These dependencies are still tested in separate test VMs, and users can still use estimators depending on them by installing the packages separately, e.g., manually, or through using the registry.deps utility.
capability tags for exogenous variables and multivariate data ( capability:exogenous , capability:multivariate ) replace the ignores-exogeneous-X and univariate-only tags, respectively. See details below.
new tags for stochasticity and determinism of estimators are introduced: capability:random_state , a boolean tag, indicating whether the estimator has a random_state parameter; and property:randomness , a string tag with values "deterministic" , "stochastic" , and "derandomized" , indicating whether the estimator is deterministic, stochastic, or deterministic after setting random_state , respectively. These tags can be used for retrieval, and are also used in the test framework. Maintainers of third party estimators should set these tags appropriately, and suite tests will check whether they are correctly set. However, code will not break if these tags are not set.
the default for capability:categorical_in_X in transformation estimators is now True . This prevents blockage in composition cases where the tag is not explicitly set, e.g., using a FunctionTransformer with a categorical compatible function in a pipeline. As this default is more permissive, user code will not break. Maintainers of third party estimators should set this tag in their transformation estimators, if they do not support categorical data in X .
a new capability:categorical_in_y tag for transformatoin estimators has been added, with a default of True . This tag indicates whether the transformation estimator can handle categorical data in y . Users whose estimators can handle y , but not categorical y , should set this tag to False . All other estimators do not need to set this tag.
The sktime.libs.pykalman module is deprecated. It will redirect to the pykalman package from 0.42.0 onwards, and will be removed entirely in version 1.0.0. Users of pykalman based estimators in sktime can continue to use them, and the estimator API itself will not change. However, users will need to install pykalman separately, as it is no longer included in sktime . Users of the sktime.libs.pykalman fork should migrate their imports to the pykalman package distributed via PyPI before sktime 1.0.0. The pykalman package is now maintained by the sktime team.
The ignores-exogeneous-X tag for forecasters is replaced by capability:exogenous . The logic of the tag is flipped, i.e., if capability:exogenous=True , the estimator can handle exogenous variables. Until 1.0.0, the old tag can still be used, but will emit a warning. Users and maintainers of third party estimators should update their code and estimators to use the new tag.
the univariate-only tag present in multiple estimator types is renamed to capability:multivariate , with boolean flip. If capability:multivariate=True , the estimator can handle multivariate data. Until 1.0.0, the old tag can still be used, but will emit a warning. Users and maintainers of third party estimators should update their code and estimators to use the new tag.
EnbPIForecaster default bootstrap method has now changed to sktime MovingBlockBootstrapTransformer
statsforecast estimators no longer imply scipy<1.16 in the python_dependencies tag, as statsforecast now supports scipy>=1.16 .
[ENH] replace ignores-exogeneous-X tag by capability:exogenous ( #8353 ) @fkiraly
[ENH] improvements to check_estimator ( #8799 ) @fkiraly
[ENH] rename univariate-only tag to capability:multivariate , with boolean flip ( #8818 ) @fkiraly
[ENH] tags for stochasticity/determinism in estimators and conditional test logic ( #8765 ) @fkiraly
[ENH] set some missing tags ( #8831 ) @fkiraly
[ENH] tag deprecation test extension to warn third party developers of deprecated tags ( #8846 ) @fkiraly
[ENH] remove dead code TagManager that was moved to scikit-base ( #8850 ) @fkiraly
[ENH] sklearn compatibility - fork scikit-learn _weighted_percentage ( #8805 ) @fkiraly
[ENH] fpp3 dataset was updated on CRAN; updated loader to use the latest ( #8778 ) @ericjb
[ENH] Dataset Loader Class for UCR UEA Datasets ( #8786 ) @jgyasu
[ENH] Dataset Loader Class for Monash Forecasting Datasets ( #8751 ) @jgyasu
[ENH] TiRex foundation model forecaster ( #8683 ) @sinemkilicdere
[ENH] skip failures of exogenous forecast tests from issue #8787 until resolved ( #8803 ) @fkiraly
[ENH] replace ignores-exogeneous-X tag by capability:exogenous ( #8353 ) @fkiraly
[ENH] Interface Toto foundation model ( #8708 ) @JATAYU000
[ENH] allow transformer instance as deseasonalizer in ThetaForecaster , fix test cases ( #8817 ) @fkiraly
[ENH] rename univariate-only tag to capability:multivariate , with boolean flip ( #8818 ) @fkiraly
[ENH] fix UpdateRefitsEvery and UpdateEvery stream update compositors to work with newer pandas DateOffset ( #8822 ) @fkiraly
[ENH] hyperactive based tuners for forecasting ( #8705 ) @fkiraly , @SimonBlanke
[ENH] add testing for PykanForecaster ( #8733 ) @fkiraly
[ENH] Update Prophetverse version bounds ( #8792 ) @felipeangelimvieira
[ENH] rename univariate-only tag to capability:multivariate , with boolean flip ( #8818 ) @fkiraly
[ENH] relax tests for TEASER estimator ( #8824 ) @fkiraly
[ENH] hyperactive based tuners for time series classification ( #8705 ) @fkiraly , @SimonBlanke
[ENH] ConvTimeNet for time series classification ( #8761 ) @Tanuj-Taneja1
[ENH] set default for capability:categorical_in_X to True for transformation estimators ( #8758 ) @fkiraly
[ENH] rename univariate-only tag to capability:multivariate , with boolean flip ( #8818 ) @fkiraly
[ENH] Element-wise arithmetic operator transformer ( #8711 ) @oresthes
[ENH] Extend IxToX with level selector shorthands and add unit tests ( #8802 ) @RobKuebler
[ENH] Add second test parameter set for SAX ( #8839 ) @piyushbiraje
[ENH] test for consistency of placeholder record signatures ( #8825 ) @fkiraly
[DOC] add missing docs page for plot_critical_difference ( #8757 ) @fkiraly
[DOC] improve docstring of check_estimator ( #8756 ) @fkiraly
[DOC] Fix Install Command ( #8788 ) @jgyasu
[DOC] better documentation of time series scitypes, full mathematical documentation of Hierarchical and Table type ( #8795 ) @fkiraly
[DOC] assorted documentation fixes ( #8816 ) @fkiraly
[DOC] pre-filled transformer extension template for series-to-features transformers ( #8834 ) @fkiraly
[DOC] restructure dataset docs ( #8854 ) @fkiraly
[DOC] Update developer guide on how to run doctests ( #8853 ) @Astrael1
[MNT] Bump actions/setup-python from 5 to 6 ( #8773 ) @dependabot[bot]
[MNT] Bump actions/setup-node from 4 to 5 ( #8772 ) @dependabot[bot]
[MNT] Update pytest-randomly requirement from <3.17,>=3.15 to >=3.15,<4.1 ( #8790 ) @dependabot[bot]
[MNT] python 3.9 end-of-life ( #8800 ) @fkiraly
[MNT] remove some rare and outdated soft dependencies ( #8804 ) @fkiraly
[MNT] remove more rare / abandoned packages from soft dependency set ( #8812 ) @fkiraly
[MNT] 0.39.0 deprecations and change actions ( #8783 ) @fkiraly
[MNT] Update ruff linting target version to python 3.10, additional drops of 3.9 references ( #8801 ) @yarnabrina
[MNT] Dependabot: Update numba requirement from <0.62 to <0.63 ( #8821 ) @dependabot[bot]
[MNT] remove python<3.13 bound in some dependency sets for numba ( #8823 ) @fkiraly
[MNT] deprecation of pykalman fork in sktime ( #8829 ) @fkiraly
[BUG] fix bug in _safe_import when importing package and pkg_name is not None ( #8763 ) @fkiraly
[BUG] Fix n_instances test and its values in the tags ( #8794 ) @jgyasu
[BUG] fix incorrect paths for UCRUEADataset ( #8815 ) @fkiraly
[BUG] fix HierarchicalToyData dataset n_instances tag value ( #8819 ) @fkiraly
[BUG] fix DirectReductionForecaster failure on newer pandas ( #8782 ) @fkiraly
[BUG] fix PykanForecaster test failures ( #8764 ) @JATAYU000
[BUG] fix predict_var of ResidualBoostingForecaster ( #8806 ) @fkiraly
[BUG] partially fix non-conformance of IGTS estimator ( #8833 ) @fkiraly
[BUG] fix DirectReductionForecaster failure on newer pandas ( #8782 ) @fkiraly
[BUG] fix MatrixProfile TypeError ( #8798 ) @JATAYU000
@Astrael1 , @ericjb , @felipeangelimvieira , @fkiraly , @JATAYU000 , @jgyasu , @oresthes , @piyushbiraje , @RobKuebler , @SimonBlanke , @sinemkilicdere , @Tanuj-Taneja1 , @tingiskhan , @yarnabrina
Release focusing on:
python 3.9 end-of-life
changes to tag framework, testing framework
scheduled deprecations
Also includes new estimators relying on changes to tag and testing framework.
TiRex foundation model forecaster (8683) sinemkilicdere
Toto foundation model forecaster (8708) JATAYU000
Dataset Loader Class to load UCR UEA Datasets from sktime huggingface repo (8786) jgyasu
Dataset Loader Class to load Monash Forecasting Datasets from sktime huggingface repo (8751) jgyasu
hyperactive based tuners for forecasting and time series classification (8705) fkiraly, SimonBlanke
prophetverse hierarchical forecasters updated; new configurable API, intermittent/hurdle distributions (8792) felipeangelimvieira, tingiskhan
ConvTimeNet for time series classification (8761) Tanuj-Taneja1
Element-wise arithmetic operator transformer (8711) oresthes
python 3.9 is no longer supported; minimum python version is now 3.10.
numba bound, where present in dependency sets, has been updated to <0.63.
the following packages have been removed from dependency sets due to being unmaintained, rarely used, or subject to strict implied secondary dependency bounds: dtaidistance, filterpy, mne, seasonal, stumpy, tbats. These dependencies are still tested in separate test VMs, and users can still use estimators depending on them by installing the packages separately, e.g., manually, or through using the registry.deps utility.
capability tags for exogenous variables and multivariate data (capability:exogenous, capability:multivariate) replace the ignores-exogeneous-X and univariate-only tags, respectively. See details below.
new tags for stochasticity and determinism of estimators are introduced: capability:random_state, a boolean tag, indicating whether the estimator has a random_state parameter; and property:randomness, a string tag with values "deterministic", "stochastic", and "derandomized", indicating whether the estimator is deterministic, stochastic, or deterministic after setting random_state, respectively. These tags can be used for retrieval, and are also used in the test framework. Maintainers of third party estimators should set these tags appropriately, and suite tests will check whether they are correctly set. However, code will not break if these tags are not set.
the default for capability:categorical_in_X in transformation estimators is now True. This prevents blockage in composition cases where the tag is not explicitly set, e.g., using a FunctionTransformer with a categorical compatible function in a pipeline. As this default is more permissive, user code will not break. Maintainers of third party estimators should set this tag in their transformation estimators, if they do not support categorical data in X.
a new capability:categorical_in_y tag for transformatoin estimators has been added, with a default of True. This tag indicates whether the transformation estimator can handle categorical data in y. Users whose estimators can handle y, but not categorical y, should set this tag to False. All other estimators do not need to set this tag.
The sktime.libs.pykalman module is deprecated. It will redirect to the pykalman package from 0.42.0 onwards, and will be removed entirely in version 1.0.0. Users of pykalman based estimators in sktime can continue to use them, and the estimator API itself will not change. However, users will need to install pykalman separately, as it is no longer included in sktime. Users of the sktime.libs.pykalman fork should migrate their imports to the pykalman package distributed via PyPI before sktime 1.0.0. The pykalman package is now maintained by the sktime team.
The ignores-exogeneous-X tag for forecasters is replaced by capability:exogenous. The logic of the tag is flipped, i.e., if capability:exogenous=True, the estimator can handle exogenous variables. Until 1.0.0, the old tag can still be used, but will emit a warning. Users and maintainers of third party estimators should update their code and estimators to use the new tag.
the univariate-only tag present in multiple estimator types is renamed to capability:multivariate, with boolean flip. If capability:multivariate=True, the estimator can handle multivariate data. Until 1.0.0, the old tag can still be used, but will emit a warning. Users and maintainers of third party estimators should update their code and estimators to use the new tag.
EnbPIForecaster default bootstrap method has now changed to sktime MovingBlockBootstrapTransformer
statsforecast estimators no longer imply scipy<1.16 in the python_dependencies tag, as statsforecast now supports scipy>=1.16.
[ENH] replace ignores-exogeneous-X tag by capability:exogenous (8353) fkiraly
[ENH] improvements to check_estimator (8799) fkiraly
[ENH] rename univariate-only tag to capability:multivariate, with boolean flip (8818) fkiraly
[ENH] tags for stochasticity/determinism in estimators and conditional test logic (8765) fkiraly
[ENH] set some missing tags (8831) fkiraly
[ENH] tag deprecation test extension to warn third party developers of deprecated tags (8846) fkiraly
[ENH] remove dead code TagManager that was moved to scikit-base (8850) fkiraly
[ENH] sklearn compatibility - fork scikit-learn _weighted_percentage (8805) fkiraly
[ENH] fpp3 dataset was updated on CRAN; updated loader to use the latest (8778) ericjb
[ENH] Dataset Loader Class for UCR UEA Datasets (8786) jgyasu
[ENH] Dataset Loader Class for Monash Forecasting Datasets (8751) jgyasu
[ENH] TiRex foundation model forecaster (8683) sinemkilicdere
[ENH] skip failures of exogenous forecast tests from issue #8787 until resolved (8803) fkiraly
[ENH] replace ignores-exogeneous-X tag by capability:exogenous (8353) fkiraly
[ENH] Interface Toto foundation model (8708) JATAYU000
[ENH] allow transformer instance as deseasonalizer in ThetaForecaster, fix test cases (8817) fkiraly
[ENH] rename univariate-only tag to capability:multivariate, with boolean flip (8818) fkiraly
[ENH] fix UpdateRefitsEvery and UpdateEvery stream update compositors to work with newer pandas DateOffset (8822) fkiraly
[ENH] hyperactive based tuners for forecasting (8705) fkiraly, SimonBlanke
[ENH] add testing for PykanForecaster (8733) fkiraly
[ENH] Update Prophetverse version bounds (8792) felipeangelimvieira
[ENH] rename univariate-only tag to capability:multivariate, with boolean flip (8818) fkiraly
[ENH] relax tests for TEASER estimator (8824) fkiraly
[ENH] hyperactive based tuners for time series classification (8705) fkiraly, SimonBlanke
[ENH] ConvTimeNet for time series classification (8761) Tanuj-Taneja1
[ENH] set default for capability:categorical_in_X to True for transformation estimators (8758) fkiraly
[ENH] rename univariate-only tag to capability:multivariate, with boolean flip (8818) fkiraly
[ENH] Element-wise arithmetic operator transformer (8711) oresthes
[ENH] Extend IxToX with level selector shorthands and add unit tests (8802) RobKuebler
[ENH] Add second test parameter set for SAX (8839) piyushbiraje
[ENH] test for consistency of placeholder record signatures (8825) fkiraly
[DOC] add missing docs page for plot_critical_difference (8757) fkiraly
[DOC] improve docstring of check_estimator (8756) fkiraly
[DOC] Fix Install Command (8788) jgyasu
[DOC] better documentation of time series scitypes, full mathematical documentation of Hierarchical and Table type (8795) fkiraly
[DOC] assorted documentation fixes (8816) fkiraly
[DOC] pre-filled transformer extension template for series-to-features transformers (8834) fkiraly
[DOC] restructure dataset docs (8854) fkiraly
[DOC] Update developer guide on how to run doctests (8853) Astrael1
[MNT] Bump actions/setup-python from 5 to 6 (8773) dependabot[bot]
[MNT] Bump actions/setup-node from 4 to 5 (8772) dependabot[bot]
[MNT] Update pytest-randomly requirement from <3.17,>=3.15 to >=3.15,<4.1 (8790) dependabot[bot]
[MNT] python 3.9 end-of-life (8800) fkiraly
[MNT] remove some rare and outdated soft dependencies (8804) fkiraly
[MNT] remove more rare / abandoned packages from soft dependency set (8812) fkiraly
[MNT] 0.39.0 deprecations and change actions (8783) fkiraly
[MNT] Update ruff linting target version to python 3.10, additional drops of 3.9 references (8801) yarnabrina
[MNT] [Dependabot](deps): Update numba requirement from <0.62 to <0.63 (8821) dependabot[bot]
[MNT] remove python<3.13 bound in some dependency sets for numba (8823) fkiraly
[MNT] deprecation of pykalman fork in sktime (8829) fkiraly
[BUG] fix bug in _safe_import when importing package and pkg_name is not None (8763) fkiraly
[BUG] Fix n_instances test and its values in the tags (8794) jgyasu
[BUG] fix incorrect paths for UCRUEADataset (8815) fkiraly
[BUG] fix HierarchicalToyData dataset n_instances tag value (8819) fkiraly
[BUG] fix DirectReductionForecaster failure on newer pandas (8782) fkiraly
[BUG] fix PykanForecaster test failures (8764) JATAYU000
[BUG] fix predict_var of ResidualBoostingForecaster (8806) fkiraly
[BUG] partially fix non-conformance of IGTS estimator (8833) fkiraly
[BUG] fix DirectReductionForecaster failure on newer pandas (8782) fkiraly
[BUG] fix MatrixProfile TypeError (8798) JATAYU000
Astrael1, ericjb, felipeangelimvieira, fkiraly, JATAYU000, jgyasu, oresthes, piyushbiraje, RobKuebler, SimonBlanke, sinemkilicdere, Tanuj-Taneja1, tingiskhan, yarnabrina
Please see our changelog for a description of all changes.
Feature release.
Please see our changelog for a description of all changes.
@aarsh1a, @aarushitandon0, @abhishek-iitmadras, @adityagarwal15, @akshathmangudi, @Alexis-Fauxbaton, @Anamika457, @Burnie-Murray, @fkiraly, @garar, @HarshvirSandhu, @hudeqiWH, @Imvedansh, @IsaiasRT, @jgyasu, @jnrusson1, @julian-fong, @LHoelper, @Mujeeb4, @OnePunchMonk, @Pradyumn-cloud, @Pranavsingh431, @Rajdeep-naha, @Sanchay117, @sinemkilicdere, @Sohaib-Ahmed21, @szepeviktor, @SzymonStolarski, @thearshkumar, @thisisrick25, @Virgile-Foussereau, @XAheli, @yarnabrina
Full Changelog: https://github.com/sktime/sktime/compare/v0.38.4...v0.38.5
Residual Boosting Forecaster, can be used to add exogenous or probabilistic capability to forecasters ( #8582 ) @Sanchay117
TinyTimeMixer forecaster now supports exogenous data ( #8709 ) @akshathmangudi
InceptionTimeClassifier now supports class weights ( #8509 ) @Virgile-Foussereau
New Dataset Downloaders for TS Datasets ( #8505 ) @jgyasu
percentage forecasting metrics now allow relative_to=y_pred ( #8335 ) @OnePunchMonk
Jarque-Bera, variance break, and Ljung-Box tests/statistics ( #8017 ) @HarshvirSandhu
temporian has been removed from the recommended transformations dependency set, as it appears to be unmaintained
esig has been removed as a soft dependency
arch (forecasting soft dependency) bounds have been updated to <7.3
Most sktime dataset loaders now use huggingface datasets as a configurable primary backend. Source files are hosted in the sktime huggingface organization.
If huggingface is not installed or available alternative backends are used, such as direct download from UCR/UEA archive.
A new tag capability:class_weight : bool has been added to indicate support for class weights. If set, a class_weight parameter can be passed to the constructor of classifiers with this tag, to specify class weights for training.
[ENH] add more common packages in standard show_versions return ( #8484 ) @fkiraly
[ENH] tests for parallelize backend abstraction utility ( #8622 ) @Sanchay117
[ENH] change MeanAbsolutePercentageError parent to BaseForecastingErrorMetric ( #8632 ) @fkiraly
[ENH] relative_to=y_pred support for percentage error forecasting metric classes ( #8335 ) @OnePunchMonk
[ENH] MASE efficient _evaluate_by_index ( #8611 ) @thearshkumar
[ENH] implement efficient _evaluate_by_index for GeometricMeanAbsoluteError ( #7095 ):user: KaustubhUp025 , @fkiraly
[ENH] Refactor Benchmarking Classes ( #8666 ) @jgyasu
[ENH] New Dataset Downloaders for TS Datasets ( #8505 ) @jgyasu
[ENH] LTSFNLinearForecaster , LTSFDLinearForecaster , and LTSFLinearForecaster get_test_params returns two sets ( #8528 ) @thearshkumar
[ENH] Added tag to PytorchForecastingTFT and removed from tests/_config ( #8559 ) @thearshkumar
[ENH] darts compatibility patch for darts 0.35.0 and higher ( #8584 ) @fkiraly
[ENH] patchtst deep learning model for forecasting ( #6787 ) @julian-fong
[ENH] Residual Booster Forecaster ( #8582 ) @Sanchay117
[ENH] split forecasting tuners module into individual files ( #8418 ) @fkiraly
[ENH] Exogenous Handling in TinyTimeMixer Forecaster ( #8709 ) @akshathmangudi
[ENH] Minor statsforecast updates ( #8692 ) @yarnabrina
[ENH] in-sample prediction capability for ARCH forecaster @SzymonStolarski
[ENH] Jarque-Bera, variance break, and Ljung-Box tests/statistics ( #8017 ) @HarshvirSandhu
[ENH] ensure that any change in classification.sklearn triggers the classification tests ( #8609 ) @fkiraly
[ENH] return histogram option in IndividualBOSS ( #8636 ) @Pradyumn-cloud
[ENH] Move FCNClassifier test skip rule to estimator tag ( #8603 ) @sinemkilicdere
[ENH] Allows 1 mismatch in RotationForest tests ( #8628 ) @akshathmangudi
[ENH] remove esig soft dependency and move esig based estimators to own test VM ( #8679 ) @fkiraly
[ENH] remove customized attentionLSTM layer. Replace with keras attention ( #8721 ) @jnrusson1
[ENH] add class weight to InceptionTimeClassifier and capabability tag ( #8509 ) @Virgile-Foussereau
[ENH] add tests for TSC evaluate parallelization backends ( #8737 ) @thisisrick25
[ENH] improve classifier evaluate defaults and documentation ( #8739 ) @fkiraly
[ENH] accept str distances in TimeSeriesDBSCAN , allow distance parameters ( #8608 ) @Burnie-Murray
[ENH] Replace deprecated _transform_single_feature with transform in _drcif_feature ( #8599 ) @Anamika457
[ENH] Second test parameter set for MiniRocketMultivariate , MiniRocketMultivariateVariable and MultiRocket ( #8631 ) @IsaiasRT
[ENH] Move WindowSummarizer test skip to estimator tag ( #8671 ) @Pranavsingh431
[ENH] remove esig soft dependency and move esig based estimators to own test VM ( #8679 ) @fkiraly
[ENH] Added tag to PytorchForecastingTFT and removed from tests/_config ( #8559 ) @thearshkumar
[ENH] tests for parallelize backend abstraction utility ( #8622 ) @Sanchay117
[DOC] improve formatting and clarity in python_dependencies tag description ( #8517 ) @fkiraly
[DOC] Add guide for module-level soft dependency imports (#7690) ( #8386 ):user: fkiraly , @aarsh1a
[DOC] fix import location of _safe_import in developer guide ( #8577 ) @fkiraly
[DOC] correct private docstrings for forecasting metrics ( #8644 ) @fkiraly
[DOC] fix some broken links on the “decision making” page ( #8583 ) @fkiraly
[DOC] add GC.OS donation mechanism to FUNDING.yml ( #8623 ) @fkiraly
[DOC] Add plot_windows to API documentation ( #8286 ) @aarushitandon0
[DOC] document metric tags in the tags API ( #8638 ) @fkiraly
[DOC] correct and improve _percentage_error docstring and related docstrings ( #8640 ) @fkiraly
[DOC] add links from estimator type specific API reference to subsection of tags API reference ( #8637 ) @fkiraly
[DOC] improved formatting and clarity in docstrings of forecasting metrics ( #8633 ) @fkiraly
[DOC] Level 2 page-toc ( #8657 ) @thearshkumar
[DOC] add ResidualBoostingForecaster to the API reference ( #8675 ) @fkiraly
[DOC] Add ClassificationBenchmark to API Reference ( #8676 ) @jgyasu
[DOC] math formula and evaluate_by_index explanation for MeanLinexError ( #8627 ) @sinemkilicdere
[DOC] improved docstrings for forecasting performance metrics ( #8682 ) @fkiraly
[DOC] Corrected typos in examples/02 classification.ipynb ( #8689 ) @hudeqiWH
[DOC] minor fixes to documentation ( #8695 ) @fkiraly
[DOC] minor formatting improvement in dependency handling guide ( #8720 ) @fkiraly
[DOC] various minor docstring formatting fixes ( #8669 ) @fkiraly
[MNT] separate dependency set for notebooks, reduce dependencies in notebook CI ( #8598 ) @fkiraly
[MNT] update notebook CI python version to 3.11 ( #8604 ) @fkiraly
[MNT] remove unmaintained (?) package temporian from recommended transformations depset ( #8485 ) @fkiraly
[MNT] reorder CI steps to minimize expected runtime ( #8557 ) @fkiraly
[MNT] Update pytest-xdist requirement from <3.8,>=3.3 to >=3.3,<3.9 ( #8470 ) @dependabot[bot]
[MNT] granular test runs in CI in single-estimator VM ( #8647 ) @fkiraly
[MNT] Fix shellcheck line ending issues on Windows ( #8502 ) @XAheli
[MNT] Bump actions/download-artifact from 4 to 5 ( #8668 ) @dependabot[bot]
[MNT] CI: change macos-13 runners to macos-latest , of arm64 Architecture ( #7715 ) @abhishek-iitmadras
[MNT] partial revert of #7715 for test-cython-estimators job. ( #8680 ) @fkiraly
[MNT] Bump actions/checkout from 4 to 5 ( #8687 ) @dependabot[bot]
[MNT] pre-commit autoupdate ( #8694 ) @yarnabrina
[MNT] Check versions in wheels workflow ( #8718 ) @szepeviktor
[MNT] ensure release depends on check_tag ( #8726 ) @fkiraly
[MNT] Exclude Python 3.13 + macos-13 combo from CI to fix torch issue (#8127) ( #8304 ) @adityagarwal15
[MNT] remove requires_cython tag and separate VM handling in favour of tests:vm ( #8674 ) @fkiraly
[MNT] Improved changelog generator to sort by subsection ( #8132 ) @Mujeeb4
[MNT] improved changelog generator: group PRs by module labels like Forecasting and Classification ( #8579 ) @sinemkilicdere
[MNT] Update test_vms to use macos-latest in CI and configure osx setup in github scripts to dynamically install dependencies ( #8624 ) @julian-fong
[MNT] Upgrade CI to set test-full to use macos-latest runners ( #8670 ) @julian-fong
[MNT] Turn off mps for macos runners on tests:vm estimators and fixes momentfm CI failures ( #8642 ) @julian-fong
[MNT] update arch bound to arch<7.3 ( #8743 ) @fkiraly
[BUG] fix ray backend unintended side effects on backend_params by replacing list.pop with list.get ( #8162 ) @LHoelper
[BUG] ensure deps dependency inspection is compatible with disjunctions in estimator dependency tags ( #8585 ) @fkiraly
[BUG] fix backend and backend_params in TSC evaluate being ignored ( #8702 ) @fkiraly
[BUG] Fix the predictions from the ARCH model implementation ( #8610 ) @SzymonStolarski
[BUG] RRF sometimes creates the wrong PeriodIndex when the frequency is higher than 1 ( #8606 ) @LHoelper
[ENH] Add full exogenous variable support for statsmodels adapter _update() method when update_params=False ( #8626 ) @Alexis-Fauxbaton
[BUG] Fix failing tests for pytorch forecasting models. ( #7844 ) @Sohaib-Ahmed21
[BUG] suppressing dict while running TTM doctest ( #8738 ) @akshathmangudi
[BUG] fix np.ndarray incompatibility in BaseDetector.fit_transform ( #8329 ) @Imvedansh
[BUG] Fix missing input conversion in BaseDetector ( #8328 ) @Rajdeep-naha
[BUG] decouple classifier, clusterer, regressor extension contract from n_jobs parameter ( #8549 ) @fkiraly
[BUG] CNTCClassifier throws error when you provide callbacks. ( #8660 ) @garar
[BUG] fix backend and backend_params in TSC evaluate being ignored ( #8702 ) @fkiraly
[BUG] fix parallelization and pickling in TSC evaluate ( #8741 ) @fkiraly
[BUG] decouple classifier, clusterer, regressor extension contract from n_jobs parameter ( #8549 ) @fkiraly
[BUG] decouple classifier, clusterer, regressor extension contract from n_jobs parameter ( #8549 ) @fkiraly
@aarsh1a , @aarushitandon0 , @abhishek-iitmadras , @adityagarwal15 , @akshathmangudi , @Alexis-Fauxbaton , @Anamika457 , @Burnie-Murray , @fkiraly , @garar , @HarshvirSandhu , @hudeqiWH , @Imvedansh , @IsaiasRT , @jgyasu , @jnrusson1 , @julian-fong , @LHoelper , @Mujeeb4 , @OnePunchMonk , @Pradyumn-cloud , @Pranavsingh431 , @Rajdeep-naha , @Sanchay117 , @sinemkilicdere , @Sohaib-Ahmed21 , @szepeviktor , @SzymonStolarski , @thearshkumar , @thisisrick25 , @Virgile-Foussereau , @XAheli , @yarnabrina
Residual Boosting Forecaster, can be used to add exogenous or probabilistic capability to forecasters (8582) Sanchay117
TinyTimeMixer forecaster now supports exogenous data (8709) akshathmangudi
InceptionTimeClassifier now supports class weights (8509) Virgile-Foussereau
New Dataset Downloaders for TS Datasets (8505) jgyasu
percentage forecasting metrics now allow relative_to=y_pred (8335) OnePunchMonk
Jarque-Bera, variance break, and Ljung-Box tests/statistics (8017) HarshvirSandhu
temporian has been removed from the recommended transformations dependency set, as it appears to be unmaintained
esig has been removed as a soft dependency
arch (forecasting soft dependency) bounds have been updated to <7.3
Most sktime dataset loaders now use huggingface datasets as a configurable primary backend. Source files are hosted in the sktime huggingface organization.
If huggingface is not installed or available alternative backends are used, such as direct download from UCR/UEA archive.
A new tag capability:class_weight : bool has been added to indicate support for class weights. If set, a class_weight parameter can be passed to the constructor of classifiers with this tag, to specify class weights for training.
[ENH] add more common packages in standard show_versions return (8484) fkiraly
[ENH] tests for parallelize backend abstraction utility (8622) Sanchay117
[ENH] change MeanAbsolutePercentageError parent to BaseForecastingErrorMetric (8632) fkiraly
[ENH] relative_to=y_pred support for percentage error forecasting metric classes (8335) OnePunchMonk
[ENH] MASE efficient _evaluate_by_index (8611) thearshkumar
[ENH] implement efficient _evaluate_by_index for GeometricMeanAbsoluteError (7095)KaustubhUp025, fkiraly
[ENH] Refactor Benchmarking Classes (8666) jgyasu
[ENH] New Dataset Downloaders for TS Datasets (8505) jgyasu
[ENH] LTSFNLinearForecaster, LTSFDLinearForecaster, and LTSFLinearForecaster get_test_params returns two sets (8528) thearshkumar
[ENH] Added tag to PytorchForecastingTFT and removed from tests/_config (8559) thearshkumar
[ENH] darts compatibility patch for darts 0.35.0 and higher (8584) fkiraly
[ENH] patchtst deep learning model for forecasting (6787) julian-fong
[ENH] Residual Booster Forecaster (8582) Sanchay117
[ENH] split forecasting tuners module into individual files (8418) fkiraly
[ENH] Exogenous Handling in TinyTimeMixer Forecaster (8709) akshathmangudi
[ENH] Minor ``statsforecast ``updates (8692) yarnabrina
[ENH] in-sample prediction capability for ARCH forecaster SzymonStolarski
[ENH] Jarque-Bera, variance break, and Ljung-Box tests/statistics (8017) HarshvirSandhu
[ENH] ensure that any change in classification.sklearn triggers the classification tests (8609) fkiraly
[ENH] return histogram option in IndividualBOSS (8636) Pradyumn-cloud
[ENH] Move FCNClassifier test skip rule to estimator tag (8603) sinemkilicdere
[ENH] Allows 1 mismatch in RotationForest tests (8628) akshathmangudi
[ENH] remove esig soft dependency and move esig based estimators to own test VM (8679) fkiraly
[ENH] remove customized attentionLSTM layer. Replace with keras attention (8721) jnrusson1
[ENH] add class weight to InceptionTimeClassifier and capabability tag (8509) Virgile-Foussereau
[ENH] add tests for TSC evaluate parallelization backends (8737) thisisrick25
[ENH] improve classifier evaluate defaults and documentation (8739) fkiraly
[ENH] accept str distances in TimeSeriesDBSCAN, allow distance parameters (8608) Burnie-Murray
[ENH] Replace deprecated _transform_single_feature with transform in _drcif_feature (8599) Anamika457
[ENH] Second test parameter set for MiniRocketMultivariate, MiniRocketMultivariateVariable and MultiRocket (8631) IsaiasRT
[ENH] Move WindowSummarizer test skip to estimator tag (8671) Pranavsingh431
[ENH] remove esig soft dependency and move esig based estimators to own test VM (8679) fkiraly
[ENH] Added tag to PytorchForecastingTFT and removed from tests/_config (8559) thearshkumar
[ENH] tests for parallelize backend abstraction utility (8622) Sanchay117
[DOC] improve formatting and clarity in python_dependencies tag description (8517) fkiraly
[DOC] Add guide for module-level soft dependency imports (#7690) (8386)fkiraly, aarsh1a
[DOC] fix import location of _safe_import in developer guide (8577) fkiraly
[DOC] correct private docstrings for forecasting metrics (8644) fkiraly
[DOC] fix some broken links on the "decision making" page (8583) fkiraly
[DOC] add GC.OS donation mechanism to FUNDING.yml (8623) fkiraly
[DOC] Add plot_windows to API documentation (8286) aarushitandon0
[DOC] document metric tags in the tags API (8638) fkiraly
[DOC] correct and improve _percentage_error docstring and related docstrings (8640) fkiraly
[DOC] add links from estimator type specific API reference to subsection of tags API reference (8637) fkiraly
[DOC] improved formatting and clarity in docstrings of forecasting metrics (8633) fkiraly
[DOC] Level 2 page-toc (8657) thearshkumar
[DOC] add ResidualBoostingForecaster to the API reference (8675) fkiraly
[DOC] Add ClassificationBenchmark to API Reference (8676) jgyasu
[DOC] math formula and evaluate_by_index explanation for MeanLinexError (8627) sinemkilicdere
[DOC] improved docstrings for forecasting performance metrics (8682) fkiraly
[DOC] Corrected typos in examples/02 classification.ipynb (8689) hudeqiWH
[DOC] minor fixes to documentation (8695) fkiraly
[DOC] minor formatting improvement in dependency handling guide (8720) fkiraly
[DOC] various minor docstring formatting fixes (8669) fkiraly
[MNT] separate dependency set for notebooks, reduce dependencies in notebook CI (8598) fkiraly
[MNT] update notebook CI python version to 3.11 (8604) fkiraly
[MNT] remove unmaintained (?) package temporian from recommended transformations depset (8485) fkiraly
[MNT] reorder CI steps to minimize expected runtime (8557) fkiraly
[MNT] Update pytest-xdist requirement from <3.8,>=3.3 to >=3.3,<3.9 (8470) dependabot[bot]
[MNT] granular test runs in CI in single-estimator VM (8647) fkiraly
[MNT] Fix shellcheck line ending issues on Windows (8502) XAheli
[MNT] Bump actions/download-artifact from 4 to 5 (8668) dependabot[bot]
[MNT] CI: change macos-13 runners to macos-latest, of arm64 Architecture (7715) abhishek-iitmadras
[MNT] partial revert of #7715 for test-cython-estimators job. (8680) fkiraly
[MNT] Bump actions/checkout from 4 to 5 (8687) dependabot[bot]
[MNT] pre-commit autoupdate (8694) yarnabrina
[MNT] Check versions in wheels workflow (8718) szepeviktor
[MNT] ensure release depends on check_tag (8726) fkiraly
[MNT] Exclude Python 3.13 + macos-13 combo from CI to fix torch issue (#8127) (8304) adityagarwal15
[MNT] remove requires_cython tag and separate VM handling in favour of tests:vm (8674) fkiraly
[MNT] Improved changelog generator to sort by subsection (8132) Mujeeb4
[MNT] improved changelog generator: group PRs by module labels like Forecasting and Classification (8579) sinemkilicdere
[MNT] Update test_vms to use macos-latest in CI and configure osx setup in github scripts to dynamically install dependencies (8624) julian-fong
[MNT] Upgrade CI to set test-full to use macos-latest runners (8670) julian-fong
[MNT] Turn off mps for macos runners on tests:vm estimators and fixes momentfm CI failures (8642) julian-fong
[MNT] update arch bound to arch<7.3 (8743) fkiraly
[BUG] fix ray backend unintended side effects on backend_params by replacing list.pop with list.get (8162) LHoelper
[BUG] ensure deps dependency inspection is compatible with disjunctions in estimator dependency tags (8585) fkiraly
[BUG] fix backend and backend_params in TSC evaluate being ignored (8702) fkiraly
[BUG] Fix the predictions from the ARCH model implementation (8610) SzymonStolarski
[BUG] RRF sometimes creates the wrong PeriodIndex when the frequency is higher than 1 (8606) LHoelper
[ENH] Add full exogenous variable support for statsmodels adapter _update() method when update_params=False (8626) Alexis-Fauxbaton
[BUG] Fix failing tests for pytorch forecasting models. (7844) Sohaib-Ahmed21
[BUG] suppressing dict while running TTM doctest (8738) akshathmangudi
[BUG] fix np.ndarray incompatibility in BaseDetector.fit_transform (8329) Imvedansh
[BUG] Fix missing input conversion in BaseDetector (8328) Rajdeep-naha
[BUG] decouple classifier, clusterer, regressor extension contract from n_jobs parameter (8549) fkiraly
[BUG] CNTCClassifier throws error when you provide callbacks. (8660) garar
[BUG] fix backend and backend_params in TSC evaluate being ignored (8702) fkiraly
[BUG] fix parallelization and pickling in TSC evaluate (8741) fkiraly
[BUG] decouple classifier, clusterer, regressor extension contract from n_jobs parameter (8549) fkiraly
[BUG] decouple classifier, clusterer, regressor extension contract from n_jobs parameter (8549) fkiraly
aarsh1a, aarushitandon0, abhishek-iitmadras, adityagarwal15, akshathmangudi, Alexis-Fauxbaton, Anamika457, Burnie-Murray, fkiraly, garar, HarshvirSandhu, hudeqiWH, Imvedansh, IsaiasRT, jgyasu, jnrusson1, julian-fong, LHoelper, Mujeeb4, OnePunchMonk, Pradyumn-cloud, Pranavsingh431, Rajdeep-naha, Sanchay117, sinemkilicdere, Sohaib-Ahmed21, szepeviktor, SzymonStolarski, thearshkumar, thisisrick25, Virgile-Foussereau, XAheli, yarnabrina
Release focusing on bugfixes, scikit-learn downwards compatibility after 1.7.0, and increasing test framework reliability.
Release focusing on bugfixes, scikit-learn downwards compatibility after 1.7.0,
and increasing test framework reliability.
For last larger feature update, see 0.37.1.
Please see our changelog for a description of all changes.
@Burnie-Murray, @ericjb, @fkiraly, @Ghxst07, @IsaiasRT, @jgyasu, @julianfong, @phoeenniixx, @thearshkumar, @Tirath5504, @xenonnn4w
Full Changelog: https://github.com/sktime/sktime/compare/v0.38.3...v0.38.4
Release focusing on bugfixes and scikit-learn downwards compatibility after 1.7.0 and increasing test framework reliability.
[ENH] The trend argument for ARDL forecaster can now also be ‘ctt’ ( #8062 ) @ericjb
[ENH] remove internal broadcasting in ChronosForecaster in favour of programmatic broadcasting ( #8561 ) @fkiraly
[ENH] more robust tag cloning in multiplexers for classification and regression ( #8550 ) @fkiraly
[ENH] move test skip config to estimators ( #8496 ) @fkiraly
[ENH] move test skips to estimators for estimators with their own VM ( #8514 ) @fkiraly
[ENH] fix some failing doctests ( #8536 ) @fkiraly
[ENH] tests:libs tag to specify library dependencies of estimators for differential testing ( #8513 ) @fkiraly
[ENH] scikit-learn and pandas downwards compatibility patch ( #8552 ) @fkiraly
[ENH] loosen condition for skpro dependency exception in BaseForecaster.predict_proba ( #8554 ) @fkiraly
[ENH] add conditional testing to transformations.hierarchical.reconcile module ( #8494 ) @fkiraly
[ENH] Move test skip config of TapNetClassifier from tests._config to estimator tags ( #8516 ) @xenonnn4w
[ENH] Move test skip config of CNTCRegressor from tests._config to estimator tags ( #8527 ) @Ghxst07
[ENH] Moved test skip config of DistanceFeatures from tests._config to estimator tags ( #8520 ) @IsaiasRT
[ENH] Moved test skip config of ResNetClassifier from tests._config to estimator tags ( #8564 ) @Burnie-Murray
[DOC] Notebook Demonstrating Benchmarking of Classifiers ( #8373 ) @jgyasu
[DOC] Correct note about unequal-length time series support in sktime ( #8510 ) @jgyasu
[DOC] fix TimeMoEForecaster doctest ( #8541 ) @fkiraly
[DOC] Update the broken all-contributors.org website link to the correct GitHub repository URL ( #8539 ) @Tirath5504
[DOC] Note about package names in _check_soft_dependencies ( #8529 ) @jgyasu
[DOC] fix broken API reference of ColumnSelect and IndexSubset ( #8504 ) @fkiraly
[DOC] A typo in test_all_estimators ( #8534 ) @phoeenniixx
[MNT] do not trigger VM tests for non-core estimators on framework changes ( #8497 ) @fkiraly
[MNT] CI steps to test by estimator with estimator specific dependencies, controlled by tag for separate VM ( #7087 ) @fkiraly , @julianfong
[MNT] fix two instances of empty strategy in GHA ( #8500 ) @fkiraly
[MNT] Dependabot: Bump actions/setup-python from 4 to 5 ( #8507 ) @dependabot[bot]
[MNT] full test run on mid-to-lower dependency version range, 2023 state of core deps ( #4016 ) @fkiraly
[MNT] replace dorny/paths-filter third party action with git diff call in detect-package-change ( #8556 ) @fkiraly
[BUG] fix sklearn tag inspection in scikit-learn < 1.6 ( #8546 ) @fkiraly
[BUG] Fix TimesFM _safe_import ( #8511 ) @thearshkumar
[BUG] fix run_test_for_module usage in tests:libs tag ( #8548 ) @fkiraly
@Burnie-Murray , @ericjb , @fkiraly , @Ghxst07 , @IsaiasRT , @jgyasu , @julianfong , @phoeenniixx , @thearshkumar , @Tirath5504 , @xenonnn4w
Release focusing on bugfixes and scikit-learn downwards compatibility after 1.7.0 and increasing test framework reliability.
[ENH] The trend argument for ARDL forecaster can now also be 'ctt' (8062) ericjb
[ENH] remove internal broadcasting in ChronosForecaster in favour of programmatic broadcasting (8561) fkiraly
[ENH] more robust tag cloning in multiplexers for classification and regression (8550) fkiraly
[ENH] move test skip config to estimators (8496) fkiraly
[ENH] move test skips to estimators for estimators with their own VM (8514) fkiraly
[ENH] fix some failing doctests (8536) fkiraly
[ENH] tests:libs tag to specify library dependencies of estimators for differential testing (8513) fkiraly
[ENH] scikit-learn and pandas downwards compatibility patch (8552) fkiraly
[ENH] loosen condition for skpro dependency exception in BaseForecaster.predict_proba (8554) fkiraly
[ENH] add conditional testing to transformations.hierarchical.reconcile module (8494) fkiraly
[ENH] Move test skip config of TapNetClassifier from tests._config to estimator tags (8516) xenonnn4w
[ENH] Move test skip config of CNTCRegressor from tests._config to estimator tags (8527) Ghxst07
[ENH] Moved test skip config of DistanceFeatures from tests._config to estimator tags (8520) IsaiasRT
[ENH] Moved test skip config of ResNetClassifier from tests._config to estimator tags (8564) Burnie-Murray
[DOC] Notebook Demonstrating Benchmarking of Classifiers (8373) jgyasu
[DOC] Correct note about unequal-length time series support in sktime (8510) jgyasu
[DOC] fix TimeMoEForecaster doctest (8541) fkiraly
[DOC] Update the broken all-contributors.org website link to the correct GitHub repository URL (8539) Tirath5504
[DOC] Note about package names in _check_soft_dependencies (8529) jgyasu
[DOC] fix broken API reference of ColumnSelect and IndexSubset (8504) fkiraly
[DOC] A typo in test_all_estimators (8534) phoeenniixx
[MNT] do not trigger VM tests for non-core estimators on framework changes (8497) fkiraly
[MNT] CI steps to test by estimator with estimator specific dependencies, controlled by tag for separate VM (7087) fkiraly, julianfong
[MNT] fix two instances of empty strategy in GHA (8500) fkiraly
[MNT] [Dependabot](deps): Bump actions/setup-python from 4 to 5 (8507) dependabot[bot]
[MNT] full test run on mid-to-lower dependency version range, 2023 state of core deps (4016) fkiraly
[MNT] replace dorny/paths-filter third party action with git diff call in detect-package-change (8556) fkiraly
[BUG] fix sklearn tag inspection in scikit-learn < 1.6 (8546) fkiraly
[BUG] Fix TimesFM _safe_import (8511) thearshkumar
[BUG] fix run_test_for_module usage in tests:libs tag (8548) fkiraly
Burnie-Murray, ericjb, fkiraly, Ghxst07, IsaiasRT, jgyasu, julianfong, phoeenniixx, thearshkumar, Tirath5504, xenonnn4w
Hotfix for ChronosForecaster and TinyTimeMixerForecaster after transformers 4.53 compatibility patch.
Hotfix for failures of ChronosForecaster and TinyTimeMixerForecaster .
[BUG] fix accidental removal of dataclass types in ChronosForecaster and TinyTimeMixerForecaster ( #8488 ) @fkiraly
transformers 4.53 compatibility patch, minor maintenance and documentation updates.
transformers 4.53 compatibility patch, minor maintenance and documentation updates.
For last larger feature update, see 0.37.1.
Please see our changelog for a description of all changes.
@abhishek-iitmadras, @fkiraly, @yarnabrina
Full Changelog: https://github.com/sktime/sktime/compare/v0.38.1...v0.38.2
transformers 4.53 compatibility patch, minor maintenance and documentation updates.
[MNT] CI: Add ubuntu-22.04-arm runners to enhance arm aarch64 architecture testing and maintenance ( #7695 ) @abhishek-iitmadras
[MNT] fix: add aarch64 installation constraint for esig and ts2vg ( #8471 ) @abhishek-iitmadras
[MNT] Skip mlflow test jobs if sktime folder or pyproject.toml is untouched ( #8474 ) @yarnabrina
[MNT] replace unsafe imports from torch and transformers with _safe_import pattern ( #8476 ) @fkiraly
[MNT] update transformers imports for 4.53 compatibility ( #8478 ) @fkiraly
[DOC] use clearer language in extension template preamble ( #8466 ) @fkiraly
[DOC] add missing API reference for GreykiteForecaster ( #8472 ) @fkiraly
@abhishek-iitmadras , @fkiraly , @yarnabrina
transformers 4.53 compatibility patch, minor maintenance and documentation updates.
[MNT] CI: Add ubuntu-22.04-arm runners to enhance arm aarch64 architecture testing and maintenance (7695) abhishek-iitmadras
[MNT] fix: add aarch64 installation constraint for esig and ts2vg (8471) abhishek-iitmadras
[MNT] Skip mlflow test jobs if sktime folder or pyproject.toml is untouched (8474) yarnabrina
[MNT] replace unsafe imports from torch and transformers with _safe_import pattern (8476) fkiraly
[MNT] update transformers imports for 4.53 compatibility (8478) fkiraly
[DOC] use clearer language in extension template preamble (8466) fkiraly
[DOC] add missing API reference for GreykiteForecaster (8472) fkiraly
abhishek-iitmadras, fkiraly, yarnabrina
scikit-learn 1.7 compatibility hotfix, thanks go to @andretheronsa.
scikit-learn 1.7 compatibility hotfix, thanks go to @andretheronsa.
For last larger feature update, see 0.37.1.
Please see our changelog for a description of all changes.
@andretheronsa, @fkiraly, @yarnabrina
Full Changelog: https://github.com/sktime/sktime/compare/v0.38.0...v0.38.1
scikit-learn 1.7 compatibility hotfix, thanks go to @andretheronsa .
[BUG] fix forecasting metric functions incompatible with scikit-learn 1.7 ( #8459 ) @andretheronsa
[MNT] revert docs to pre-8385 state ( #8454 , #8444 ) @fkiraly , @yarnabrina
@andretheronsa , @fkiraly , @yarnabrina
scikit-learn 1.7 compatibility hotfix, thanks go to andretheronsa.
[BUG] fix forecasting metric functions incompatible with scikit-learn 1.7 (8459) andretheronsa
[MNT] revert docs to pre-8385 state (8454, 8444) fkiraly, yarnabrina
andretheronsa, fkiraly, yarnabrina
Maintenance release with scheduled deprecations and change actions.
Maintenance release with scheduled deprecations and change actions.
For last larger feature update, see 0.37.1.
Please see our changelog for a description of all changes.
@felipeangelimvieira, @fkiraly, @jgyasu, @yarnabrina
Full Changelog: https://github.com/sktime/sktime/compare/v0.37.1...v0.38.0
Maintenance release with scheduled deprecations and change actions.
For the last non-maintenance content update, see 0.37.1.
numpy (core dependency) bounds have been updated to <2.4
pandas (core dependency) bounds have been updated to <2.4
scikit-learn (core dependency) bounds have been updated to <1.8
scipy (core dependency) bounds have been updated to <1.17
deprecated methods from BaseBenchmark have been removed
forecasting_validation utility has been removed - users should call ForecastingBenchmark instead
unused copies datasets have been removed from sktime.datasets.data . Users should replace any direct read of data files in sktime.datasets.data with loading via dataset classes or loader functions. For example, reading ArrowHead_TRAIN.tsv should be replaced by calling load from sktime.datasets.classification.ArrowHead .
removed: the proba module has been removed from sktime . Users should replace imports from sktime.proba with imports from skpro.distributions , in the skpro package.
[MNT] Dependabot: Update pandas requirement from <2.3.0,>=1.1 to >=1.1,<2.4.0 ( #8349 ) @dependabot[bot]
[MNT] lint conf.py ( #8437 ) @fkiraly
[MNT] increase joblib bound to joblib<1.7 ( #8424 ) @fkiraly
[MNT] revert documentation PR 8385 ( #8442 ) @fkiraly
[MNT] add scipy<1.16 bounds to incompatible estimators ( #8447 ) @fkiraly
[MNT] raise numpy bound to numpy<2.4 ( #8414 ) @fkiraly
[MNT] 0.38.0 deprecations and change actions ( #8429 ) @fkiraly
[MNT] Delete Redundant Dataset Files ( #8441 ) @jgyasu
[MNT] sklearn 1.7 compatibility ( #8439 ) @fkiraly
[MNT] raise scikit-learn bound to <1.8.0 ( #8402 ) @fkiraly
[MNT] Delete Redundant Dataset Files ( #8441 ) @jgyasu
[DOC] Add new reconcilers to API reference ( #8440 ) @felipeangelimvieira
[BUG] Documentation conf.py is missing mandatory extensions ( #8435 ) @felipeangelimvieira
[BUG] fix sphinx changelog link generation and panel overview ( #8432 ) @fkiraly
@felipeangelimvieira , @fkiraly , @jgyasu , @yarnabrina
Maintenance release with scheduled deprecations and change actions.
For the last non-maintenance content update, see 0.37.1.
numpy (core dependency) bounds have been updated to <2.4
pandas (core dependency) bounds have been updated to <2.4
scikit-learn (core dependency) bounds have been updated to <1.8
scipy (core dependency) bounds have been updated to <1.17
deprecated methods from BaseBenchmark have been removed
forecasting_validation utility has been removed - users should call ForecastingBenchmark instead
unused copies datasets have been removed from sktime.datasets.data. Users should replace any direct read of data files in sktime.datasets.data with loading via dataset classes or loader functions. For example, reading ArrowHead_TRAIN.tsv should be replaced by calling load from sktime.datasets.classification.ArrowHead.
removed: the proba module has been removed from sktime. Users should replace imports from sktime.proba with imports from skpro.distributions, in the skpro package.
[MNT] [Dependabot](deps): Update pandas requirement from <2.3.0,>=1.1 to >=1.1,<2.4.0 (8349) dependabot[bot]
[MNT] lint conf.py (8437) fkiraly
[MNT] increase joblib bound to joblib<1.7 (8424) fkiraly
[MNT] revert documentation PR 8385 (8442) fkiraly
[MNT] add scipy<1.16 bounds to incompatible estimators (8447) fkiraly
[MNT] raise numpy bound to numpy<2.4 (8414) fkiraly
[MNT] 0.38.0 deprecations and change actions (8429) fkiraly
[MNT] Delete Redundant Dataset Files (8441) jgyasu
[MNT] sklearn 1.7 compatibility (8439) fkiraly
[MNT] raise scikit-learn bound to <1.8.0 (8402) fkiraly
[MNT] Delete Redundant Dataset Files (8441) jgyasu
[DOC] Add new reconcilers to API reference (8440) felipeangelimvieira
[BUG] Documentation conf.py is missing mandatory extensions (8435) felipeangelimvieira
[BUG] fix sphinx changelog link generation and panel overview (8432) fkiraly
felipeangelimvieira, fkiraly, jgyasu, yarnabrina
Please see our changelog for a description of all changes.
Feature release.
Please see our changelog for a description of all changes.
@amanmogal, @andoriyaprashant, @Ankit-1204, @benHeid, @felipeangelimvieira, @fkiraly, @gbilleyPeco, @HarshvirSandhu, @jgyasu, @jobs-git, @julian-fong, @ksharma6, @marrov, @mateuszkasprowicz, @mohamed-halemo, @nahcol10, @nmwitzig, @OmBiradar, @pcpp94, @piyushbiraje, @Prahitha, @Pranavsingh431, @RobKuebler, @szepeviktor, @Tveten, @vedantag17, @yury-fedotov
Full Changelog: https://github.com/sktime/sktime/compare/v0.37.0...v0.37.1
new reconciler module, new non-negative reconciliation ( #7697 ) @felipeangelimvieira
ClassificationBenchmark and evaluate utility for benchmarking time series classifiers ( #8267 , #8357 ) @ksharma6 , @jgyasu
momentfm foundation model forecaster ( #6570 ) @julian-fong
momentfm foundation model time series classifier ( #8263 ) @julian-fong
ForecastKnownValues can now handle multi-index prescribed forecasts ( #8236 ) @RobKuebler
sliding greedy splitter ( #8168 ) @marrov
forecast evaluate utility can now handle temporal and global splits ( #8064 ) @benHeid
SkforecastRecursive now supports in-sample predict_quantiles ( #8255 ) @nahcol10
optuna (tuning soft dependency) bounds have been updated to <4.5
[ENH] improve formatting of _safe_import docstrings ( #8226 ) @fkiraly
[ENH] _placeholder_record to allow mirroring of multiple locations, update of skchange import locations ( #8232 ) @fkiraly
[ENH] avoid all-tags retrieval, add package name to tag aliaser ( #8401 ) @fkiraly
[ENH] Add temporal global split to forecast evaluate utility ( #8064 ) @benHeid
[ENH] sliding greedy splitter ( #8168 ) @marrov
[ENH] make numpy output coercions consistent in metrics and distances ( #8194 ) @fkiraly
[ENH] Visualisation utility for global splits ( #8204 ) @benHeid
[ENH] Add relative_to parameter to percentage error metrics to enable relative-by-prediction metrics - part 1, functions ( #8266 ) @OmBiradar
[ENH] refactor forecasting metrics into individual private submodules ( #8279 ) @fkiraly
[ENH] enforce MeanSquaredErrorPercentage to be positive ( #8318 ) @fkiraly
[ENH] Use sktime Provided Dataset in TSC Dataset Loaders ( #8416 ) @jgyasu
[ENH] Add is_hierarchical method to check whether a dataframe is hierarchical ( #8170 ) @gbilleyPeco
[ENH] Speed up docs building and minor docs fixes ( #8385 ) @mateuszkasprowicz
[ENH] interfacing momentfm forecaster as a deep learning module ( #6570 ) @julian-fong
[ENH] make ForecastKnownValues multiindex-ready ( #8236 ) @RobKuebler
[ENH] Add regularization hyperparameter in ReconcilerForecaster ( #7660 ) @HarshvirSandhu
[ENH] Greykite forecaster integration ( #7902 ) @vedantag17
[ENH] change forecaster default tag value of capability:categorical_in_X to True to enable composition case ( #8231 ) @fkiraly
[ENH] SkforecastRecursive - support in-sample predict_quantiles ( #8255 ) @nahcol10
[ENH] enable univariate mode for DynamicFactor forecaster ( #8345 ) @fkiraly
[ENH] minor improvements to MultiplexForecaster ( #8422 ) @fkiraly
[ENH] Fix MLflow 3.0 compatibility while maintaining 2.x support ( #8395 ) @amanmogal
[ENH] skchange homogenization: variable identification + minor fixes ( #7509 ) @Tveten
[ENH] update of skchange import locations ( #8232 ) @fkiraly
[ENH] Extending categorical framework support in X to time series classifiers ( #8114 ) @Ankit-1204
[ENH] Make grid search parameters in ShapeDTW configurable to enable faster testing ( #8324 ) @benHeid
[ENH] momentfm classifier ( #8263 ) @julian-fong
[ENH] evaluate for TSC and TSR ( #8267 , #6888 ) @jgyasu , @ksharma6
[ENH] Second Set of Test Parameters TimeSeriesForest ( #8337 ) @nmwitzig
[ENH] Add predict_proba tag for classifier to estimator overview table ( #8344 ) @jgyasu
[ENH] Add parallelization support for TSC evaluate ( #8347 ) @jgyasu
[ENH] ClassificationBenchmark for benchmarking time series classifiers ( #8357 ) @jgyasu
[ENH] Tests for TSC evaluate Parallel Backend ( #8359 ) @jgyasu
[ENH] make numpy output coercions consistent in metrics and distances ( #8194 ) @fkiraly
[ENH] Extending categorical framework support in X to time series regressors ( #8114 ) @Ankit-1204
[ENH] evaluate for TSC and TSR ( #8267 , #6888 ) @jgyasu , @ksharma6
[ENH] Enable EnbPIForecaster to work with sktime native bootstrap algorithms ( #7562 ) @benHeid
[ENH] Simplify HampelFilter code and add a unit test ( #8249 ) @RobKuebler
[ENH] Add missing space to Pipeline constructor warning ( #8275 ) @yury-fedotov
[ENH] Implement reconcilers as transformations, and add new non-negative reconciliation ( #7697 ) @felipeangelimvieira
[ENH] faster MovingBlockBootstrapTransformer tests, improved docstring ( #8420 ) @fkiraly
[ENH] move run_doctest to scikit-base ( #8319 ) @fkiraly
[ENH] refactor test flag logic to tests._config ( #8000 ) @fkiraly
[ENH] test framework: “core classes” and non-core classes with different test trigger conditions ( #8408 ) @fkiraly
[ENH] core dependency changes also trigger estimator tests ( #8415 ) @fkiraly
[ENH] remove some unstable and long running estimators from the core testing set ( #8421 ) @fkiraly
[DOC] mathematical reference for probabilistic forecasting metrics ( #8213 ) @fkiraly
[DOC] improve VARReduce docstring ( #8179 ) @fkiraly
[DOC] GC.OS sponsoring badge in README.md ( #8320 ) @fkiraly
[DOC] add missing imports in fracdiff docstrings ( #8317 ) @fkiraly
[DOC] fix broken docstring of KNeighborsTimeSeriesClassifierTslearn ( #8341 ) @fkiraly
[DOC] mathematical formulae for MASE and MedASE ( #8339 ) @piyushbiraje
[DOC] Add reference to the updated roadmap page in contributing.rst ( #8350 ) @jgyasu
[DOC] fix KNeighborsTimeSeriesClassifierTslearn docstring ( #8410 ) @fkiraly
[DOC] Fix citation formatting in lower_bounding.py to resolve doc warnings ( #8389 ) @mohamed-halemo
[DOC] Fix failing doctest in ARLagOrderSelector example ( #8407 ) @andoriyaprashant
[DOC] fix MomentFMClassifier docstring example ( #8419 ) @fkiraly
[DOC] Update glossary.rst ( #8214 ) @pcpp94
[DOC] Improves documentation in Testing and continuous integration webpage ( #8247 ) @julian-fong
[DOC] Fix Docstring Example of Distance Based Classifiers ( #8313 ) @jgyasu
[DOC] Update utils.rst API reference to add plot_windows ( #7993 ) @Prahitha
[DOC] Fix documentation typos throughout the code base5 ( #8301 ) @szepeviktor
[MNT] remove soft dependency from ForecastingOptunaSearchCV docstring example ( #8202 ) @fkiraly
[MNT] fix VECM docstring example skip ( #8207 ) @fkiraly
[MNT] fix apply_split doctest and skip SARIMAX failures ( #8220 ) @fkiraly
[MNT] remove stefanzweifel/git-auto-commit-action from all-contributors workflow ( #8369 ) @fkiraly
[MNT] add scipy<1.16 bound to statsforecast and pmdarima based forecasters ( #8427 ) @fkiraly
[MNT] Dependabot: Update pytest-timeout requirement from <2.4,>=2.1 to >=2.1,<2.5 ( #8242 ) @dependabot[bot]
[MNT] Dependabot: Update pytest-xdist requirement from <3.7,>=3.3 to >=3.3,<3.8 ( #8303 ) @dependabot[bot]
[MNT] Dependabot: Update pytest requirement from <8.4,>=7.4 to >=7.4,<8.5 ( #8338 ) @dependabot[bot]
[MNT] Dependabot: Update pytorch-forecasting requirement from <1.4.0,>=1.0.0 to >=1.0.0,<1.5.0 ( #8399 ) @dependabot[bot]
[MNT] Dependabot: Update optuna requirement from <4.3 to <4.4 ( #8197 ) @dependabot[bot]
[MNT] Dependabot: Update optuna requirement from <4.4 to <4.5 ( #8400 ) @dependabot[bot]
[BUG] fix links to documentation from estimator html repr ( #8178 ) @mateuszkasprowicz
[BUG] _safe_import to avoid failures due to soft dependency chains in sktime/libs/uni2ts/forecast.py ( #8192 ) @jgyasu
[BUG] fix _check_soft_dependencies error message if version is wrong ( #8430 ) @fkiraly
[BUG] Fix fh handling in greedy splitters ( #8200 ) @marrov
[BUG] Fix "kkwargs" typo in ForecastingBenchmark ( #8306 ) @yury-fedotov
[BUG] Fix boundary equality in EmpiricalCoverage ( #8372 ) @marrov
[BUG] Fix aggregation function assignment in EnbPI class ( #8230 ) @marrov
[BUG] fix recursive reduction tests ( #8274 ) @fkiraly
[BUG] fix pandas future warning in pytorch-forecasting adapter due to in-place modification of a copy ( #8269 ) @jobs-git
[BUG] fix: use names keyword instead of name ( #8257 ) @RobKuebler
[BUG] fix wrong return of fit in doc tests of MoiraiForecaster and TimeLLM ( #8326 ) @benHeid
[BUG] fix incorrect scitype inference in forecasting pipelines for polymorphic forecasters ( #8237 ) @nahcol10
[BUG] remove broken reducer tests ( #8412 ) @fkiraly
[BUG] Fixed AlignerDTW when window_size needs to be passed given window_type ( #8124 ) @OmBiradar
[BUG] fix RotationForest for a custom base_estimator ( #8270 ) @fkiraly
[BUG] reindex features in TSFreshClassifier.predict_proba to match training columns ( #8285 ) @Pranavsingh431
[BUG] Fix Aggregator inverse_transform being skipped ( #8358 ) @felipeangelimvieira
[BUG] fix accidental failure in fixture generating utility _get_parallel_test_fixtures , ( #8193 ) @fkiraly
[BUG] fix differential testing logic for DateTimeFeatures specific test ( #8293 ) @andoriyaprashant
@amanmogal , @andoriyaprashant , @Ankit-1204 , @benHeid , @felipeangelimvieira , @fkiraly , @gbilleyPeco , @HarshvirSandhu , @jgyasu , @jobs-git , @julian-fong , @ksharma6 , @marrov , @mateuszkasprowicz , @mohamed-halemo , @nahcol10 , @nmwitzig , @OmBiradar , @pcpp94 , @piyushbiraje , @Prahitha , @Pranavsingh431 , @RobKuebler , @szepeviktor , @Tveten , @vedantag17 , @yury-fedotov
new reconciler module, new non-negative reconciliation (7697) felipeangelimvieira
ClassificationBenchmark and evaluate utility for benchmarking time series classifiers (8267, 8357) ksharma6, jgyasu
momentfm foundation model forecaster (6570) julian-fong
momentfm foundation model time series classifier (8263) julian-fong
ForecastKnownValues can now handle multi-index prescribed forecasts (8236) RobKuebler
sliding greedy splitter (8168) marrov
forecast evaluate utility can now handle temporal and global splits (8064) benHeid
SkforecastRecursive now supports in-sample predict_quantiles (8255) nahcol10
optuna (tuning soft dependency) bounds have been updated to <4.5
[ENH] improve formatting of _safe_import docstrings (8226) fkiraly
[ENH] _placeholder_record to allow mirroring of multiple locations, update of skchange import locations (8232) fkiraly
[ENH] avoid all-tags retrieval, add package name to tag aliaser (8401) fkiraly
[ENH] Add temporal global split to forecast evaluate utility (8064) benHeid
[ENH] sliding greedy splitter (8168) marrov
[ENH] make numpy output coercions consistent in metrics and distances (8194) fkiraly
[ENH] Visualisation utility for global splits (8204) benHeid
[ENH] Add relative_to parameter to percentage error metrics to enable relative-by-prediction metrics - part 1, functions (8266) OmBiradar
[ENH] refactor forecasting metrics into individual private submodules (8279) fkiraly
[ENH] enforce MeanSquaredErrorPercentage to be positive (8318) fkiraly
[ENH] Use sktime Provided Dataset in TSC Dataset Loaders (8416) jgyasu
[ENH] Add is_hierarchical method to check whether a dataframe is hierarchical (8170) gbilleyPeco
[ENH] Speed up docs building and minor docs fixes (8385) mateuszkasprowicz
[ENH] interfacing momentfm forecaster as a deep learning module (6570) julian-fong
[ENH] make ForecastKnownValues multiindex-ready (8236) RobKuebler
[ENH] Add regularization hyperparameter in ReconcilerForecaster (7660) HarshvirSandhu
[ENH] Greykite forecaster integration (7902) vedantag17
[ENH] change forecaster default tag value of capability:categorical_in_X to True to enable composition case (8231) fkiraly
[ENH] SkforecastRecursive - support in-sample predict_quantiles (8255) nahcol10
[ENH] enable univariate mode for DynamicFactor forecaster (8345) fkiraly
[ENH] minor improvements to MultiplexForecaster (8422) fkiraly
[ENH] Fix MLflow 3.0 compatibility while maintaining 2.x support (8395) amanmogal
[ENH] skchange homogenization: variable identification + minor fixes (7509) Tveten
[ENH] update of skchange import locations (8232) fkiraly
[ENH] Extending categorical framework support in X to time series classifiers (8114) Ankit-1204
[ENH] Make grid search parameters in ShapeDTW configurable to enable faster testing (8324) benHeid
[ENH] momentfm classifier (8263) julian-fong
[ENH] evaluate for TSC and TSR (8267, 6888) jgyasu, ksharma6
[ENH] Second Set of Test Parameters TimeSeriesForest (8337) nmwitzig
[ENH] Add predict_proba tag for classifier to estimator overview table (8344) jgyasu
[ENH] Add parallelization support for TSC evaluate (8347) jgyasu
[ENH] ClassificationBenchmark for benchmarking time series classifiers (8357) jgyasu
[ENH] Tests for TSC evaluate Parallel Backend (8359) jgyasu
[ENH] make numpy output coercions consistent in metrics and distances (8194) fkiraly
[ENH] Extending categorical framework support in X to time series regressors (8114) Ankit-1204
[ENH] evaluate for TSC and TSR (8267, 6888) jgyasu, ksharma6
[ENH] Enable EnbPIForecaster to work with sktime native bootstrap algorithms (7562) benHeid
[ENH] Simplify HampelFilter code and add a unit test (8249) RobKuebler
[ENH] Add missing space to Pipeline constructor warning (8275) yury-fedotov
[ENH] Implement reconcilers as transformations, and add new non-negative reconciliation (7697) felipeangelimvieira
[ENH] faster MovingBlockBootstrapTransformer tests, improved docstring (8420) fkiraly
[ENH] move run_doctest to scikit-base (8319) fkiraly
[ENH] refactor test flag logic to tests._config (8000) fkiraly
[ENH] test framework: "core classes" and non-core classes with different test trigger conditions (8408) fkiraly
[ENH] core dependency changes also trigger estimator tests (8415) fkiraly
[ENH] remove some unstable and long running estimators from the core testing set (8421) fkiraly
[DOC] mathematical reference for probabilistic forecasting metrics (8213) fkiraly
[DOC] improve VARReduce docstring (8179) fkiraly
[DOC] GC.OS sponsoring badge in README.md (8320) fkiraly
[DOC] add missing imports in fracdiff docstrings (8317) fkiraly
[DOC] fix broken docstring of KNeighborsTimeSeriesClassifierTslearn (8341) fkiraly
[DOC] mathematical formulae for MASE and MedASE (8339) piyushbiraje
[DOC] Add reference to the updated roadmap page in contributing.rst (8350) jgyasu
[DOC] fix KNeighborsTimeSeriesClassifierTslearn docstring (8410) fkiraly
[DOC] Fix citation formatting in lower_bounding.py to resolve doc warnings (8389) mohamed-halemo
[DOC] Fix failing doctest in ARLagOrderSelector example (8407) andoriyaprashant
[DOC] fix MomentFMClassifier docstring example (8419) fkiraly
[DOC] Update glossary.rst (8214) pcpp94
[DOC] Improves documentation in Testing and continuous integration webpage (8247) julian-fong
[DOC] Fix Docstring Example of Distance Based Classifiers (8313) jgyasu
[DOC] Update utils.rst API reference to add plot_windows (7993) Prahitha
[DOC] Fix documentation typos throughout the code base5 (8301) szepeviktor
[MNT] remove soft dependency from ForecastingOptunaSearchCV docstring example (8202) fkiraly
[MNT] fix VECM docstring example skip (8207) fkiraly
[MNT] fix apply_split doctest and skip SARIMAX failures (8220) fkiraly
[MNT] remove stefanzweifel/git-auto-commit-action from all-contributors workflow (8369) fkiraly
[MNT] add scipy<1.16 bound to statsforecast and pmdarima based forecasters (8427) fkiraly
[MNT] [Dependabot](deps): Update pytest-timeout requirement from <2.4,>=2.1 to >=2.1,<2.5 (8242) dependabot[bot]
[MNT] [Dependabot](deps): Update pytest-xdist requirement from <3.7,>=3.3 to >=3.3,<3.8 (8303) dependabot[bot]
[MNT] [Dependabot](deps): Update pytest requirement from <8.4,>=7.4 to >=7.4,<8.5 (8338) dependabot[bot]
[MNT] [Dependabot](deps): Update pytorch-forecasting requirement from <1.4.0,>=1.0.0 to >=1.0.0,<1.5.0 (8399) dependabot[bot]
[MNT] [Dependabot](deps): Update optuna requirement from <4.3 to <4.4 (8197) dependabot[bot]
[MNT] [Dependabot](deps): Update optuna requirement from <4.4 to <4.5 (8400) dependabot[bot]
[BUG] fix links to documentation from estimator html repr (8178) mateuszkasprowicz
[BUG] _safe_import to avoid failures due to soft dependency chains in sktime/libs/uni2ts/forecast.py (8192) jgyasu
[BUG] fix _check_soft_dependencies error message if version is wrong (8430) fkiraly
[BUG] Fix fh handling in greedy splitters (8200) marrov
[BUG] Fix "kkwargs" typo in ForecastingBenchmark (8306) yury-fedotov
[BUG] Fix boundary equality in EmpiricalCoverage (8372) marrov
[BUG] Fix aggregation function assignment in EnbPI class (8230) marrov
[BUG] fix recursive reduction tests (8274) fkiraly
[BUG] fix pandas future warning in pytorch-forecasting adapter due to in-place modification of a copy (8269) jobs-git
[BUG] fix: use names keyword instead of name (8257) RobKuebler
[BUG] fix wrong return of fit in doc tests of MoiraiForecaster and TimeLLM (8326) benHeid
[BUG] fix incorrect scitype inference in forecasting pipelines for polymorphic forecasters (8237) nahcol10
[BUG] remove broken reducer tests (8412) fkiraly
[BUG] Fixed AlignerDTW when window_size needs to be passed given window_type (8124) OmBiradar
[BUG] fix RotationForest for a custom base_estimator (8270) fkiraly
[BUG] reindex features in TSFreshClassifier.predict_proba to match training columns (8285) Pranavsingh431
[BUG] Fix Aggregator inverse_transform being skipped (8358) felipeangelimvieira
[BUG] fix accidental failure in fixture generating utility _get_parallel_test_fixtures, (8193) fkiraly
[BUG] fix differential testing logic for DateTimeFeatures specific test (8293) andoriyaprashant
amanmogal, andoriyaprashant, Ankit-1204, benHeid, felipeangelimvieira, fkiraly, gbilleyPeco, HarshvirSandhu, jgyasu, jobs-git, julian-fong, ksharma6, marrov, mateuszkasprowicz, mohamed-halemo, nahcol10, nmwitzig, OmBiradar, pcpp94, piyushbiraje, Prahitha, Pranavsingh431, RobKuebler, szepeviktor, Tveten, vedantag17, yury-fedotov
Maintenance release with scheduled deprecations and change actions.
Maintenance release with scheduled deprecations and change actions.
For the last non-maintenance content update, see 0.36.1.
keras-self-attention (deep learning soft dependency) is no longer a soft dependency. The package has been abandoned and is now vendored in sktime .
Forecasting and transformation pipelines can now be built with an scikit-learn estimator directly, without needing to wrap in TabularToSeriesAdaptor .
The behaviour is the same as before, but the TabularToSeriesAdaptor is now optional.
deprecated: the handles-missing-data tag is deprecated and renamed to capability:missing_values , to be consistent with other estimator types. The old tag name handles-missing-data can be used until 1.0.0, when it will be removed, but will raise a deprecation warning.
removed: the deprecated HierarchyEnsembleForecaster.fitted_list parameter has been removed. Users should use get_fitted_params or forecasters_ instead.
removed: the deprecated annotation module has been removed. Users should replace imports from sktime.annotation with imports from sktime.detection .
removed: the deprecated Y argument in detector methods has been removed. Instead, users should use the y argument.
[ENH] _safe_import to allow multiple inheritance from multiple mock classes ( #8061 ) @fkiraly
[ENH] refactor datatypes registry to scikit-base records ( #7816 ) @fkiraly
[ENH] remove accidental univariate-only tags from forecasters ( #8180 ) @fkiraly
[ENH] start rename of handles-missing-data tag to capability:missing_values ( #7917 ) @fkiraly
[ENH] automatic transformer coercion in explicit forecasting pipeline specification syntax ( #8153 ) @fkiraly
[ENH] automatic transformer coercion in explicit transformer pipeline specification syntax ( #8160 ) @fkiraly
[MNT] 0.37.0 deprecations and change actions ( #8148 ) @fkiraly
[MNT] skip some known failing tests ( #8152 ) @fkiraly
[MNT] docstring example fixes and test skips ( #8159 ) @fkiraly
[MNT] temporary skip SCINetForecaster ( #8171 ) @fkiraly
[MNT] remove tsbootstrap soft dependency from all_extras_pandas2 dependency set ( #8174 ) @fkiraly
[MNT] vendor abandoned keras-self-attention library ( #8155 ) @fkiraly
[MNT] skip tests with ray parallelization fixtures ( #8188 ) @fkiraly
[MNT] skip VECM doctest due to pandas 1 incompatibility ( #8187 ) @fkiraly
[BUG] fix scitype return in case coerce_to_list=False and force_single_scitype=False if two or more types are detected ( #8133 ) @fkiraly
[BUG] Stop Importing Modules if Raised an Exception in _safe_import ( #8182 ) @jgyasu
@fkiraly , @jgyasu
Maintenance release with scheduled deprecations and change actions.
For the last non-maintenance content update, see 0.36.1.
keras-self-attention (deep learning soft dependency) is no longer a soft dependency. The package has been abandoned and is now vendored in sktime.
Forecasting and transformation pipelines can now be built with an scikit-learn estimator directly, without needing to wrap in TabularToSeriesAdaptor.
The behaviour is the same as before, but the TabularToSeriesAdaptor is now optional.
deprecated: the handles-missing-data tag is deprecated and renamed to capability:missing_values, to be consistent with other estimator types. The old tag name handles-missing-data can be used until 1.0.0, when it will be removed, but will raise a deprecation warning.
removed: the deprecated HierarchyEnsembleForecaster.fitted_list parameter has been removed. Users should use get_fitted_params or forecasters_ instead.
removed: the deprecated annotation module has been removed. Users should replace imports from sktime.annotation with imports from sktime.detection.
removed: the deprecated Y argument in detector methods has been removed. Instead, users should use the y argument.
[ENH] _safe_import to allow multiple inheritance from multiple mock classes (8061) fkiraly
[ENH] refactor datatypes registry to scikit-base records (7816) fkiraly
[ENH] remove accidental univariate-only tags from forecasters (8180) fkiraly
[ENH] start rename of handles-missing-data tag to capability:missing_values (7917) fkiraly
[ENH] automatic transformer coercion in explicit forecasting pipeline specification syntax (8153) fkiraly
[ENH] automatic transformer coercion in explicit transformer pipeline specification syntax (8160) fkiraly
[MNT] 0.37.0 deprecations and change actions (8148) fkiraly
[MNT] skip some known failing tests (8152) fkiraly
[MNT] docstring example fixes and test skips (8159) fkiraly
[MNT] temporary skip SCINetForecaster (8171) fkiraly
[MNT] remove tsbootstrap soft dependency from all_extras_pandas2 dependency set (8174) fkiraly
[MNT] vendor abandoned keras-self-attention library (8155) fkiraly
[MNT] skip tests with ray parallelization fixtures (8188) fkiraly
[MNT] skip VECM doctest due to pandas 1 incompatibility (8187) fkiraly
[BUG] fix scitype return in case coerce_to_list=False and force_single_scitype=False if two or more types are detected (8133) fkiraly
[BUG] Stop Importing Modules if Raised an Exception in _safe_import (8182) jgyasu
fkiraly, jgyasu
Please see our changelog for a description of all changes.
Feature release.
Please see our changelog for a description of all changes.
@amitsubhashchejara, @Ankit-1204, @benHeid, @Bergschaf, @celestinoxp, @danferns, @ericjb, @felipeangelimvieira, @fkiraly, @geetu040, @haroon0x, @hazrulakmal, @itsbharatj, @jgyasu, @LHoelper, @LorchZachery, @oseiskar, @phoeenniixx, @PranavBhatP, @Reckadon, @satvshr, @Sohaib-Ahmed21, @Spinachboul, @tapyu, @TomatoChocolate12, @Utkarsh-Aggarwal, @vedantag17, @wilsbj, @XinyuWuu, @yarnabrina
Full Changelog: https://github.com/sktime/sktime/compare/v0.36.0...v0.36.1
Time-LLM forecaster ( #7663 ) @jgyasu
TimeMoE forecaster ( #7869 ) @PranavBhatP
support for chronos-bolt in ChronosForecaster ( #7718 ) @PranavBhatP
simdkalman -based Kalman filter transformers ( #8030 ) @oseiskar
ray parallelization backend ( #8031 ) @LHoelper
HierarchyEnsembleForecaster can now select different parallelization backends ( #7954 ) @LHolper
ADIDA intermittency forecaster from statsforecast ( #7754 ) @vedantag17
SeasonalDummies - one hot encoding for seasonal categorical variable transformer ( #7915 ) @ericjb
Revised Forecasting Benchmarking Module ( #7603 ) @benHeid
tensorflow (deep learning soft dependency) bounds have been updated to <2.20
dask (data container and parallelization back-end soft dependency) bounds have been updated to <2025.2.1,>2024.8.2
ray is now a soft dependency (parallelization back-end)
all native parallelization locations now accept "ray" as a parallelization backend, for example forecasting parameter tuning, or hierarchical parallelization via set_config
ForecastingBenchmark has been reworked:
accepts sktime -in-memory data formats and dataset loaders
provides different handlers for benchmarking outputs, including json , csv , and parquet .
supports global forecasters
[ENH] Add utility _safe_import for programmatic soft dependency isolation ( #7702 ) @jgyasu
[ENH] private utilities for safe argument passing to function with unknown signature ( #7766 ) @fkiraly
[ENH] extend _check_soft_dependencies to allow dependency disjunctions ( #7943 ) @fkiraly
[ENH] refactor scitype detection logic and meta-estimator clone logic to include sklearn estimator types ( #8131 ) @fkiraly
[ENH] Extend forecasting benchmark to global setting ( #6774 ) @XinyuWuu
[ENH] Add error handling to cINNForecaster ’s CurveFitForecaster ( #7602 ) @Utkarsh-Aggarwal
[ENH] Revision Forecasting Benchmarking Module ( #7603 ) @benHeid
[ENH] Extend benchmarking to accept sktime-in-memory data formats ( #5055 ) @hazrulakmal
[ENH] benchmarking: make write to csv file optional ( #7850 ) @fkiraly
[ENH] speed up forecasting evaluate benchmark input checks ( #7899 ) @fkiraly
[ENH] coerce scalar forecasting metric outputs to float ( #8074 ) @fkiraly
[ENH] Ensure Forecasting Metrics Return np.float64 ( #8099 ) @amitsubhashchejara
[ENH] Data loader functions for datasets to return y with numeric dtype ( #7588 ) @vedantag17
[ENH] STLForecaster - added method plot_components ( #6423 ) @ericjb
[ENH] ADIDA intermittency forecaster from statsforecast ( #7754 ) @vedantag17
[ENH] add clusterer use to GroupByCategoryForecaster and docstrings ( #7782 ) @fkiraly
[ENH] Implement support for chronos-bolt in ChronosForecaster ( #7718 ) @PranavBhatP
[ENH] HierarchyEnsembleForecaster parallelization backends ( #7954 ) @LHoelper
[ENH] improve methodology test coverage for DirectReductionForecaster and RecursiveReductionForecaster ( #7770 ) @PranavBhatP
[ENH] reducer - minor code clarity refactor ( #7556 ) @fkiraly
[ENH] add fit_strategy for TinyTimeMixerForecaster ( #7648 ) @Sohaib-Ahmed21
[ENH] TimeLLM for Forecasting (vendored) ( #7663 ) @jgyasu
[ENH] Enable the HF Transformers Forecaster to pass model as object ( #7689 ) @Spinachboul
[ENH] ES-RNN Forecaster ( #7772 ) @Ankit-1204
[ENH] move naive forecasters to separate folder ( #8067 ) @fkiraly
[ENH] TimeMoE forecaster (vendored) ( #7869 ) @PranavBhatP
[ENH] Extended parameter estimator base class with y in fit ( #7737 ) @satvshr
[ENH] test parameter estimator update in suite tests ( #7952 ) @fkiraly
[ENH] change imports utility to produce public import paths ( #8073 ) @fkiraly
[ENH] Test case to ensure an AssertionError is thrown if EXCLUDED_TESTS_BY_TEST is not updated ( #7622 ) @satvshr
[ENH] fix test_excluded_tests_by_test for HFTransformersForecaster ( #7882 ) @fkiraly
[ENH] improve condition and error message in test_excluded_tests_by_test ( #7904 ) @fkiraly
[ENH] trigger differential tests if core framework modules change ( #7923 ) @fkiraly
[ENH] add some missing differential test skips ( #8004 ) @fkiraly
[ENH] check_estimator verbose output on passed and failed tests if raise_exceptions=True ( #8047 ) @fkiraly
[ENH] handle BaseForecaster.predict_proba skpro dependency in TestAllEstimators ( #8023 ) @fkiraly
[ENH] minor linting fix in tests._config ( #7912 ) @fkiraly
[ENH] Relax Panel-To-Panel transformations to allow changing the number of instances ( #7991 ) @felipeangelimvieira
[ENH] catch test_fit_idempotent predict_proba soft dependency error ( #8076 ) @fkiraly
[ENH] Time Series AUPRC Anomaly detection metric based on QuoVadisTAD ( #7729 ) @Ankit-1024
[ENH] param sets for MiniRocket ( #7814 ) @Bergschaf
[ENH] added test parameters to Rocket estimator ( #7813 ) @fkiraly
[ENH] Adds Seasonal dummies features via a transformer SeasonalDummies (one hot encoding for seasonal categorical variable) ( #7915 ) @ericjb
[ENH] repeater bootstrap ( #5854 ) @fkiraly
[ENH] Use RBFLayer to do the transformations in RBFTransformer and move RBFLayer and RBFNetwork to networks ( #7730 ) @phoeenniixx
[ENH] Use valid covariance matrices in Kalman filter unit tests ( #8029 ) @oseiskar
[ENH] move Kalman filter transformers into folder module ( #8066 ) @fkiraly
[ENH] Add simdkalman -based Series and Panel transformers ( #8030 ) @oseiskar
[ENH] Allow Merger to work on hierarchical inputs ( #8071 ) @wilsbj
[DOC] minor fixes to data types documentation ( #7777 ) @fkiraly
[DOC] fix link to developer guide in new contributor instructions ( #7842 ) @fkiraly , @yarnabrina
[DOC] update enbpi docstrings ( #7835 ) @Spinachboul
[DOC] Creating template for performance metrics for detection ( #7740 ) @Ankit-1204
[DOC] fixing examples link ( #7911 ) @fkiraly
[DOC] FCNClassifier - add Wang et al reference ( #7884 ) @itsbharatj
[DOC] added docstring to the HierarchicalDask class ( #7684 ) @TomatoChocolate12
[DOC] clarify assumptions on function fun in utils.parallel.parallelize ( #7823 ) @fkiraly
[DOC] minor typo fix in add_estimators.rst ( #7956 ) @fkiraly
[DOC] Added example for AlignerNaive and AlignerLuckyDtw ( #7795 ) @haroon0x
[DOC] fix RotationForest docstring example ( #7978 ) @fkiraly
[DOC] minor linting improvement ( #7987 ) @fkiraly
[DOC] extension templates improvements ( #7898 ) @fkiraly
[DOC] improved API reference for python_dependencies tag ( #8022 ) @fkiraly
[DOC] fix FCNClassifier docstring ( #8044 ) @tapyu
[DOC] fix a few broken glossary refs ( #8050 ) @fkiraly
[DOC] add clear reference to ReconcilerForecaster in Reconciler documentation ( #8045 ) @fkiraly
[DOC] remove some failing docstring examples ( #8121 ) @fkiraly
[DOC] docstring example fixes - functions ( #8126 ) @fkiraly
[MNT] Removes MiniRocket from EXCLUDED_TESTS_BY_TEST list to fix CI failures ( #7860 ) @satvshr
[MNT] Add LLM prompt to template ( #7851 ) @felipeangelimvieira
[MNT] Avoid skipping of unit tests of several forecasters ( #7826 ) @yarnabrina
[MNT] temporary skip of MACNNClassifier failing tests ( #7959 ) @fkiraly
[MNT] temporary skip of MCDCNNClassifier failing tests ( #7960 ) @fkiraly
[MNT] remove tj-actions ( #7989 ) @fkiraly
[MNT] move release to trusted publishers ( #7990 ) @fkiraly
[MNT] temporary skip of VARMAX tests due to timeout ( #7986 ) @fkiraly
[MNT] isolate pykan imports inside estimator ( #6885 ) @fkiraly
[MNT] temporary skip of pytorch-forecasting tests until resolution of #7997 ( #7998 ) @fkiraly
[MNT] Updating pre-commit hooks and running all hooks on all files ( #7996 ) @yarnabrina
[MNT] temporary skip of STDBSCAN clusterer ( #8024 ) @fkiraly
[MNT] For TSFreshFeatureExtractor remove upper bound of scipy set lower for tsfresh ( #7847 ) @benHeid
[MNT] lint test_kalman_filter ( #8054 ) @fkiraly
[MNT] increase pytorch-forecasting bound ( #7783 ) @fkiraly
[MNT] Update pyproject.toml to install tensorflow up until python 3.12 ( #7865 ) @Reckadon
[MNT] remove numpy 2 dependency handler from BaseObject ( #7955 ) @fkiraly
[MNT] address numpy array assignment deprecation ( #7919 ) @fkiraly
[MNT] address some pandas deprecation warnings ( #7866 ) @fkiraly
[MNT] temporary skip for DistanceFeatures until fix of known failures ( #8078 ) @fkiraly
[MNT] temporary skip for TimeSeriesKvisibility until fix of known failures ( #8079 ) @fkiraly
[MNT] Removes MiniRocket from EXCLUDED_TESTS_BY_TEST list to fix CI failures ( #7860 ) @satvshr
[MNT] temporary skip of pytorch-forecasting tests until resolution of #7997 ( #7998 ) @fkiraly
[MNT] Temporarily remove RecursiveReductionForecaster from unit tests ( #8081 ) @felipeangelimvieira
[MNT] skip early classifier TEASER in test_multiprocessing_idempotent ( #8085 ) @fkiraly
[MNT] skip test_transformation_can_return_new_instances until resolution of #8084 ( #8086 ) @fkiraly
[MNT] remove dask from 3.13 dep sets due to failure of dependency resolution ( #8101 ) @fkiraly
[MNT] Remove debugging line from test_check_estimator ( #8104 ) @felipeangelimvieira
[MNT] update PyKANForecaster dependency set ( #8109 ) @fkiraly
[MNT] remove CNTCRegressor parameter test skip ( #8122 ) @fkiraly
[MNT] prevent torch installation on python 3.13 and macos ( #8143 ) @fkiraly
[MNT] increase tensorflow bound to <2.20 ( #8119 ) @fkiraly
[MNT] update dask version to dask<2025.2.1,>2024.8.2 ( #8083 ) @fkiraly
[MNT] Dependabot: Update sphinx-gallery requirement from <0.19.0 to <0.20.0 ( #7819 ) @dependabot[bot]
[MNT] Dependabot: Update dask requirement from <2024.8.1 to <2025.2.1 ( #7857 ) @dependabot[bot]
[BUG] Fix _safe_import and Add Tests ( #7888 ) @jgyasu
[BUG] safer context manager for pandas 2.2 warning suppression ( #7910 ) @fkiraly
[BUG] fix _check_estimator_deps in disjunction case ( #8025 ) @fkiraly
[BUG] Add missing frequencies to load_forecastingdata ( #7575 ) @benHeid
[BUG] fix load_fpp3 with fallback URL to avoid failure from changing CRAN URLs ( #7799 ) @ericjb
[BUG] conversion Series - pd.DataFrame to Series - pd.Series now retains original column name as attr name in pd.Series ( #7762 ) @ericjb
[BUG] fix unpack error from check_align in datatypes.alignment ( #7821 ) @geetu040
[BUG] fix BaggingForecaster.predict_proba ( #7791 ) @fkiraly
[BUG] fix (str, forecaster, count) parametrization for EnsembleForecaster ( #7785 ) @fkiraly
[BUG] fix statsmodels adapter variable name for pd.Series ( #7832 ) @fkiraly
[BUG] test_get_test_params_coverage throws an error when all tests are triggered ( #7878 ) @satvshr
[BUG] Fix RecursiveReductionForecaster failing test_hierarchical_with_exogeneous ( #7927 ) @felipeangelimvieira
[BUG] hack-fix for round trip conversions of pandas time series ( #7925 ) @fkiraly
[BUG] ensure statsforecast adapter keeps track of trained forecaster ( #7983 ) @yarnabrina
[BUG] Fix HierarchyEnsembleForecaster test failure ( #7992 ) @wilsbj
[BUG] fix _HeterogenousEnsembleForecaster parameter inspection ( #7984 ) @fkiraly
[BUG] fix HeterogenousMetaForecaster handling of forecasters ( #8020 ) @fkiraly
[BUG] Use _DelegatedForecaster in HierarchicalProphet , add upper bound to prophetverse version ( #7980 ) @felipeangelimvieira
[BUG] MAPA forecaster - missing clone, using exogenous data ( #7872 ) @satvshr
[BUG] fix timestamp format change in fh.to_absolute ( #7914 , #8110 ) @XinyuWuu
[BUG] fix freq loss in NaiveForecaster ( #8091 ) @XinyuWuu
[BUG] fix parameter estimator missing univariate scenario and masked bugs ( #7932 ) @fkiraly
[BUG] fix failing TestAllForecasters test collection due to pytest inheritance bug ( #7789 ) @yarnabrina
[BUG] fix check_estimator integration for TestAllGlobalForecasters ( #7800 ) @fkiraly , @sohaib-ahmed21
[BUG] in tests, handle soft dependency in BaseForecaster.predict_proba on skpro ( #7981 ) @fkiraly
[BUG] Testing util retrieve_scenarios should select scenarios for forecasters, clusterers, and classifiers ( #7971 ) @wilsbj
[BUG] catch test_fit_idempotent predict_proba soft dependency error - secondary fix ( #8080 ) @fkiraly
[BUG] ensure test_doctest_examples checks doctests correctly ( #8065 ) @fkiraly
[BUG] BaseDetector - Incorrect docstring and tests for segments_to_change_points ( #7640 ) @satvshr
[BUG] fix sporadic out-of-bounds error in ProximityStump and WDTW ( #7836 ) @fkiraly
[BUG] skip RocketClassifier expected output test until resolved ( #7922 ) @fkiraly
[BUG] Fix serialization errors in tests of MACNNClassifier ( #8057 ) @benHeid
[BUG] TimeSeriesKVisibility - fix missing python_dependencies tag ( #7972 ) @fkiraly
[BUG] fix broken clustering tests in TestAllClusterers ( #7977 ) @fkiraly
[BUG] Fix failing tests for SklearnClustererPipeline , TimeSeriesDBSCAN , ClustererPipeline ( #8028 ) @wilsbj
[BUG] squeeze output of CNTCRegressor ( #8116 ) @benHeid
[BUG] sklearn 1.6.X compatibility for TabularToSeriesAdaptor - handle stateless tag by requires_fit in adapt.py ( #7845 ) @benHeid
[BUG] Fix infinite recursion in _SuppressWarningPattern and LossySetitemError in SummaryTransformer ( #7903 ) @celestinoxp
[BUG] fix Series -to- Primitives transformer output index for hierarchical data ( #8037 ) @fkiraly
[BUG] Handle hierarchical mtype in ClustererAsTransformer ( #8013 ) @felipeangelimvieira
[BUG] Hierarchical data passed to ClustererAsTransformer fails ( #8053 ) @wilsbj
[BUG] fix expected output length tests for hierarchical transformations ( #8069 ) @fkiraly
[BUG] Make _TransformChangeNInstances capable of handling hierarchical test scenarios ( #8092 ) @felipeangelimvieira
[BUG] Handle hierarchical input types in ColumnConcatenator correctly ( #8070 ) @wilsbj
[BUG] Fix VMD odd length data returns even length decomposition ( #8094 ) @danferns , @LorchZachery
@amitsubhashchejara , @Ankit-1204 , @benHeid , @Bergschaf , @celestinoxp , @danferns , @ericjb , @felipeangelimvieira , @fkiraly , @geetu040 , @haroon0x , @hazrulakmal , @itsbharatj , @jgyasu , @LHoelper , @LorchZachery , @oseiskar , @phoeenniixx , @PranavBhatP , @Reckadon , @satvshr , @Sohaib-Ahmed21 , @Spinachboul , @tapyu , @TomatoChocolate12 , @Utkarsh-Aggarwal , @vedantag17 , @wilsbj , @XinyuWuu , @yarnabrina
Time-LLM forecaster (7663) jgyasu
TimeMoE forecaster (7869) PranavBhatP
support for chronos-bolt in ChronosForecaster (7718) PranavBhatP
simdkalman-based Kalman filter transformers (8030) oseiskar
ray parallelization backend (8031) LHoelper
HierarchyEnsembleForecaster can now select different parallelization backends (7954) LHolper
ADIDA intermittency forecaster from statsforecast (7754) vedantag17
SeasonalDummies - one hot encoding for seasonal categorical variable transformer (7915) ericjb
Revised Forecasting Benchmarking Module (7603) benHeid
tensorflow (deep learning soft dependency) bounds have been updated to <2.20
dask (data container and parallelization back-end soft dependency) bounds have been updated to <2025.2.1,>2024.8.2
ray is now a soft dependency (parallelization back-end)
forecasting parameter tuning, or hierarchical parallelization via set_config
ForecastingBenchmark has been reworked:
accepts sktime-in-memory data formats and dataset loaders
provides different handlers for benchmarking outputs, including json, csv, and parquet.
supports global forecasters
[ENH] Add utility _safe_import for programmatic soft dependency isolation (7702) jgyasu
[ENH] private utilities for safe argument passing to function with unknown signature (7766) fkiraly
[ENH] extend _check_soft_dependencies to allow dependency disjunctions (7943) fkiraly
[ENH] refactor scitype detection logic and meta-estimator clone logic to include sklearn estimator types (8131) fkiraly
[ENH] Extend forecasting benchmark to global setting (6774) XinyuWuu
[ENH] Add error handling to cINNForecaster's CurveFitForecaster (7602) Utkarsh-Aggarwal
[ENH] Revision Forecasting Benchmarking Module (7603) benHeid
[ENH] Extend benchmarking to accept sktime-in-memory data formats (5055) hazrulakmal
[ENH] benchmarking: make write to csv file optional (7850) fkiraly
[ENH] speed up forecasting evaluate benchmark input checks (7899) fkiraly
[ENH] coerce scalar forecasting metric outputs to float (8074) fkiraly
[ENH] Ensure Forecasting Metrics Return np.float64 (8099) amitsubhashchejara
[ENH] Data loader functions for datasets to return y with numeric dtype (7588) vedantag17
[ENH] STLForecaster - added method plot_components (6423) ericjb
[ENH] ADIDA intermittency forecaster from statsforecast (7754) vedantag17
[ENH] add clusterer use to GroupByCategoryForecaster and docstrings (7782) fkiraly
[ENH] Implement support for chronos-bolt in ChronosForecaster (7718) PranavBhatP
[ENH] HierarchyEnsembleForecaster parallelization backends (7954) LHoelper
[ENH] improve methodology test coverage for DirectReductionForecaster and RecursiveReductionForecaster (7770) PranavBhatP
[ENH] reducer - minor code clarity refactor (7556) fkiraly
[ENH] add fit_strategy for TinyTimeMixerForecaster (7648) Sohaib-Ahmed21
[ENH] TimeLLM for Forecasting (vendored) (7663) jgyasu
[ENH] Enable the HF Transformers Forecaster to pass model as object (7689) Spinachboul
[ENH] ES-RNN Forecaster (7772) Ankit-1204
[ENH] move naive forecasters to separate folder (8067) fkiraly
[ENH] TimeMoE forecaster (vendored) (7869) PranavBhatP
[ENH] Extended parameter estimator base class with y in fit (7737) satvshr
[ENH] test parameter estimator update in suite tests (7952) fkiraly
[ENH] change imports utility to produce public import paths (8073) fkiraly
[ENH] Test case to ensure an AssertionError is thrown if EXCLUDED_TESTS_BY_TEST is not updated (7622) satvshr
[ENH] fix test_excluded_tests_by_test for HFTransformersForecaster (7882) fkiraly
[ENH] improve condition and error message in test_excluded_tests_by_test (7904) fkiraly
[ENH] trigger differential tests if core framework modules change (7923) fkiraly
[ENH] add some missing differential test skips (8004) fkiraly
[ENH] check_estimator verbose output on passed and failed tests if raise_exceptions=True (8047) fkiraly
[ENH] handle BaseForecaster.predict_proba skpro dependency in TestAllEstimators (8023) fkiraly
[ENH] minor linting fix in tests._config (7912) fkiraly
[ENH] Relax Panel-To-Panel transformations to allow changing the number of instances (7991) felipeangelimvieira
[ENH] catch test_fit_idempotent predict_proba soft dependency error (8076) fkiraly
[ENH] Time Series AUPRC Anomaly detection metric based on QuoVadisTAD (7729) Ankit-1024
[ENH] param sets for MiniRocket (7814) Bergschaf
[ENH] added test parameters to Rocket estimator (7813) fkiraly
[ENH] Adds Seasonal dummies features via a transformer SeasonalDummies (one hot encoding for seasonal categorical variable) (7915) ericjb
[ENH] repeater bootstrap (5854) fkiraly
[ENH] Use RBFLayer to do the transformations in RBFTransformer and move RBFLayer and RBFNetwork to networks (7730) phoeenniixx
[ENH] Use valid covariance matrices in Kalman filter unit tests (8029) oseiskar
[ENH] move Kalman filter transformers into folder module (8066) fkiraly
[ENH] Add simdkalman-based Series and Panel transformers (8030) oseiskar
[ENH] Allow Merger to work on hierarchical inputs (8071) wilsbj
[DOC] minor fixes to data types documentation (7777) fkiraly
[DOC] fix link to developer guide in new contributor instructions (7842) fkiraly, yarnabrina
[DOC] update enbpi docstrings (7835) Spinachboul
[DOC] Creating template for performance metrics for detection (7740) Ankit-1204
[DOC] fixing examples link (7911) fkiraly
[DOC] FCNClassifier - add Wang et al reference (7884) itsbharatj
[DOC] added docstring to the HierarchicalDask class (7684) TomatoChocolate12
[DOC] clarify assumptions on function fun in utils.parallel.parallelize (7823) fkiraly
[DOC] minor typo fix in add_estimators.rst (7956) fkiraly
[DOC] Added example for AlignerNaive and AlignerLuckyDtw (7795) haroon0x
[DOC] fix RotationForest docstring example (7978) fkiraly
[DOC] minor linting improvement (7987) fkiraly
[DOC] extension templates improvements (7898) fkiraly
[DOC] improved API reference for python_dependencies tag (8022) fkiraly
[DOC] fix FCNClassifier docstring (8044) tapyu
[DOC] fix a few broken glossary refs (8050) fkiraly
[DOC] add clear reference to ReconcilerForecaster in Reconciler documentation (8045) fkiraly
[DOC] remove some failing docstring examples (8121) fkiraly
[DOC] docstring example fixes - functions (8126) fkiraly
[MNT] Removes MiniRocket from EXCLUDED_TESTS_BY_TEST list to fix CI failures (7860) satvshr
[MNT] Add LLM prompt to template (7851) felipeangelimvieira
[MNT] Avoid skipping of unit tests of several forecasters (7826) yarnabrina
[MNT] temporary skip of MACNNClassifier failing tests (7959) fkiraly
[MNT] temporary skip of MCDCNNClassifier failing tests (7960) fkiraly
[MNT] remove tj-actions (7989) fkiraly
[MNT] move release to trusted publishers (7990) fkiraly
[MNT] temporary skip of VARMAX tests due to timeout (7986) fkiraly
[MNT] isolate pykan imports inside estimator (6885) fkiraly
[MNT] temporary skip of pytorch-forecasting tests until resolution of #7997 (7998) fkiraly
[MNT] Updating pre-commit hooks and running all hooks on all files (7996) yarnabrina
[MNT] temporary skip of STDBSCAN clusterer (8024) fkiraly
[MNT] For TSFreshFeatureExtractor remove upper bound of scipy set lower for tsfresh (7847) benHeid
[MNT] lint test_kalman_filter (8054) fkiraly
[MNT] increase pytorch-forecasting bound (7783) fkiraly
[MNT] Update pyproject.toml to install tensorflow up until python 3.12 (7865) Reckadon
[MNT] remove numpy 2 dependency handler from BaseObject (7955) fkiraly
[MNT] address numpy array assignment deprecation (7919) fkiraly
[MNT] address some pandas deprecation warnings (7866) fkiraly
[MNT] temporary skip for DistanceFeatures until fix of known failures (8078) fkiraly
[MNT] temporary skip for TimeSeriesKvisibility until fix of known failures (8079) fkiraly
[MNT] Removes MiniRocket from EXCLUDED_TESTS_BY_TEST list to fix CI failures (7860) satvshr
[MNT] temporary skip of pytorch-forecasting tests until resolution of #7997 (7998) fkiraly
[MNT] Temporarily remove RecursiveReductionForecaster from unit tests (8081) felipeangelimvieira
[MNT] skip early classifier TEASER in test_multiprocessing_idempotent (8085) fkiraly
[MNT] skip test_transformation_can_return_new_instances until resolution of #8084 (8086) fkiraly
[MNT] remove dask from 3.13 dep sets due to failure of dependency resolution (8101) fkiraly
[MNT] Remove debugging line from test_check_estimator (8104) felipeangelimvieira
[MNT] update PyKANForecaster dependency set (8109) fkiraly
[MNT] remove CNTCRegressor parameter test skip (8122) fkiraly
[MNT] prevent torch installation on python 3.13 and macos (8143) fkiraly
[MNT] increase tensorflow bound to <2.20 (8119) fkiraly
[MNT] update dask version to dask<2025.2.1,>2024.8.2 (8083) fkiraly
[MNT] [Dependabot](deps): Update sphinx-gallery requirement from <0.19.0 to <0.20.0 (7819) dependabot[bot]
[MNT] [Dependabot](deps): Update dask requirement from <2024.8.1 to <2025.2.1 (7857) dependabot[bot]
[BUG] Fix _safe_import and Add Tests (7888) jgyasu
[BUG] safer context manager for pandas 2.2 warning suppression (7910) fkiraly
[BUG] fix _check_estimator_deps in disjunction case (8025) fkiraly
[BUG] Add missing frequencies to load_forecastingdata (7575) benHeid
[BUG] fix load_fpp3 with fallback URL to avoid failure from changing CRAN URLs (7799) ericjb
[BUG] conversion Series-pd.DataFrame to Series-pd.Series now retains original column name as attr name in pd.Series (7762) ericjb
[BUG] fix unpack error from check_align in datatypes.alignment (7821) geetu040
[BUG] fix BaggingForecaster.predict_proba (7791) fkiraly
[BUG] fix (str, forecaster, count) parametrization for EnsembleForecaster (7785) fkiraly
[BUG] fix statsmodels adapter variable name for pd.Series (7832) fkiraly
[BUG] test_get_test_params_coverage throws an error when all tests are triggered (7878) satvshr
[BUG] Fix RecursiveReductionForecaster failing test_hierarchical_with_exogeneous (7927) felipeangelimvieira
[BUG] hack-fix for round trip conversions of pandas time series (7925) fkiraly
[BUG] ensure statsforecast adapter keeps track of trained forecaster (7983) yarnabrina
[BUG] Fix HierarchyEnsembleForecaster test failure (7992) wilsbj
[BUG] fix _HeterogenousEnsembleForecaster parameter inspection (7984) fkiraly
[BUG] fix _HeterogenousMetaForecaster handling of forecasters_ (8020) fkiraly
[BUG] Use _DelegatedForecaster in HierarchicalProphet, add upper bound to prophetverse version (7980) felipeangelimvieira
[BUG] MAPA forecaster - missing clone, using exogenous data (7872) satvshr
[BUG] fix timestamp format change in fh.to_absolute (7914, 8110) XinyuWuu
[BUG] fix freq loss in NaiveForecaster (8091) XinyuWuu
[BUG] fix parameter estimator missing univariate scenario and masked bugs (7932) fkiraly
[BUG] fix failing TestAllForecasters test collection due to pytest inheritance bug (7789) yarnabrina
[BUG] fix check_estimator integration for TestAllGlobalForecasters (7800) fkiraly, sohaib-ahmed21
[BUG] in tests, handle soft dependency in BaseForecaster.predict_proba on skpro (7981) fkiraly
[BUG] Testing util retrieve_scenarios should select scenarios for forecasters, clusterers, and classifiers (7971) wilsbj
[BUG] catch test_fit_idempotent predict_proba soft dependency error - secondary fix (8080) fkiraly
[BUG] ensure test_doctest_examples checks doctests correctly (8065) fkiraly
[BUG] BaseDetector - Incorrect docstring and tests for segments_to_change_points (7640) satvshr
[BUG] fix sporadic out-of-bounds error in ProximityStump and WDTW (7836) fkiraly
[BUG] skip RocketClassifier expected output test until resolved (7922) fkiraly
[BUG] Fix serialization errors in tests of MACNNClassifier (8057) benHeid
[BUG] TimeSeriesKVisibility - fix missing python_dependencies tag (7972) fkiraly
[BUG] fix broken clustering tests in TestAllClusterers (7977) fkiraly
[BUG] Fix failing tests for SklearnClustererPipeline, TimeSeriesDBSCAN, ClustererPipeline (8028) wilsbj
[BUG] squeeze output of CNTCRegressor (8116) benHeid
[BUG] sklearn 1.6.X compatibility for TabularToSeriesAdaptor - handle stateless tag by requires_fit in adapt.py (7845) benHeid
[BUG] Fix infinite recursion in _SuppressWarningPattern and LossySetitemError in SummaryTransformer (7903) celestinoxp
[BUG] fix Series-to-Primitives transformer output index for hierarchical data (8037) fkiraly
[BUG] Handle hierarchical mtype in ClustererAsTransformer (8013) felipeangelimvieira
[BUG] Hierarchical data passed to ClustererAsTransformer fails (8053) wilsbj
[BUG] fix expected output length tests for hierarchical transformations (8069) fkiraly
[BUG] Make _TransformChangeNInstances capable of handling hierarchical test scenarios (8092) felipeangelimvieira
[BUG] Handle hierarchical input types in ColumnConcatenator correctly (8070) wilsbj
[BUG] Fix VMD odd length data returns even length decomposition (8094) danferns, LorchZachery
amitsubhashchejara, Ankit-1204, benHeid, Bergschaf, celestinoxp, danferns, ericjb, felipeangelimvieira, fkiraly, geetu040, haroon0x, hazrulakmal, itsbharatj, jgyasu, LHoelper, LorchZachery, oseiskar, phoeenniixx, PranavBhatP, Reckadon, satvshr, Sohaib-Ahmed21, Spinachboul, tapyu, TomatoChocolate12, Utkarsh-Aggarwal, vedantag17, wilsbj, XinyuWuu, yarnabrina
Maintenance release with scikit-learn 1.6.X and numpy 2.2.X compatibility.
Maintenance release with scikit-learn 1.6.X and numpy 2.2.X compatibility.
For the last non-maintenance content update, see 0.35.1.
Full Changelog: https://github.com/sktime/sktime/compare/v0.35.1...v0.36.0
Maintenance release with scheduled updates to version bounds.
For the last non-maintenance content update, see 0.35.1.
numpy (core dependency) bounds have been updated to <2.3.0
scikit-learn (core dependency) bounds have been updated to <1.7.0
[ENH] sklearn 1.6.X compatibility patch ( #7751 , #7755 ) @fkiraly
[ENH] fix minor typos in error messages ( #7750 ) @fkiraly
[MNT] Dependabot: Update numpy requirement from <2.2,>=1.21 to >=1.21,<2. ( #7675 , #7747 ) @dependabot[bot]
[MNT] raise scikit-learn bound to <1.7.0 ( #7742 ) @fkiraly
[MNT] fix some faulty instances of dependency handling ( #7760 ) @fkiraly
[BUG] fix BaseDatatype.call dispatching to private, not public method ( #7748 ) @fkiraly
Maintenance release with scheduled updates to version bounds.
For the last non-maintenance content update, see 0.35.1.
numpy (core dependency) bounds have been updated to <2.3.0
scikit-learn (core dependency) bounds have been updated to <1.7.0
[ENH] sklearn 1.6.X compatibility patch (7751, 7755) fkiraly
[ENH] fix minor typos in error messages (7750) fkiraly
[MNT] [Dependabot](deps): Update numpy requirement from <2.2,>=1.21 to >=1.21,<2. (7675, 7747) dependabot[bot]
[MNT] raise scikit-learn bound to <1.7.0 (7742) fkiraly
[MNT] fix some faulty instances of dependency handling (7760) fkiraly
[BUG] fix BaseDatatype.__call__ dispatching to private, not public method (7748) fkiraly
Please see our changelog for a description of all changes.
Feature release.
Please see our changelog for a description of all changes.
@Abelarm, @abhishek-iitmadras, @Adarsh2345, @Alex-JG3, @Ankit-1204, @b9junkers, @benHeid, @cheachu, @Dehelaan, @felipeangelimvieira, @fkiraly, @fnhirwa, @gavinkatz001, @gbilleyPeco, @geetu040, @HarshvirSandhu, @jgyasu, @keitaVigano, @KrishBakshi, @ksharma6, @lenaklosik, @marcosfelt, @marrov, @mateuszkasprowicz, @phoeenniixx, @PranavBhatP, @RHYTHM2405, @RUPESH-KUMAR01, @SABARNO-PRAMANICK, @Salzemann, @sanskarmodi8, @satvshr, @seigpe, @skinan, @Spinachboul, @tanvincible, @VjayRam, @y-mx, @yarnabrina
Full Changelog: https://github.com/sktime/sktime/compare/v0.35.0...v0.35.1
Forecasting metrics call can be passed by_index to return per-index metric evaluate_by_index ( #7608 ) @benHeid
MAPA forecaster ( #7620 ) @phoeenniixx
Interface to skforecast.recursive.ForecasterRecursive ( #7554 ) @yarnabrina
Native reducers (2nd generation) now support arbitrary imputation strategies ( #7535 , #7646 ) @lenaklosik
New detection metrics: RAND Index, windowed F1-score ( #7533 , #7628 ) @gavinkatz001
forecasting evaluate now allows returning fitted estimators via return_model parameter ( #7637 ) @marrov
Tags and object API for Datasets ( #7398 ) @felipeangelimvieira
K-visibility clustering algorithm ( #7592 ) @seigpe
Interface to statmodels ar_select_order for lag order estimation ( #7693 ) @satvshr
skpro (probability distributions soft dependency) bounds have been updated to >=2,<2.10.0
numba (computation soft dependency) bounds have been updated to <0.62
optuna (hyperparameter optimization soft dependency) bounds have been updated to <4.3
pykan (forecasting soft dependency) bounds have been updated to >=0.2.1,<0.2.9
dask (data container and parallelization back-end soft dependency) bounds have been updated to <2025.1.1
a unified API has been introduced for data sets and data loaders,
following the scikit-base pattern. A unified load method is now available for all data loaders, with further unified API points such as keys and tags for retrieval. For further details, see Datasets reference.
All Forecasting metrics can now be constructed with a by_index=True argument.
This will result in a direct call returning the metric per time index. This is equivalent to a call of the evaluate_by_index method. by_index=False is the default, and dispatches, as currently, to the evaluate method of the metric.
In HierarchyEnsembleForecaster , the fitted_list attribute is deprecated. To access the fitted estimators, users should use instead either the get_fitted_params method, or the attribute forecasters_ . Given a fitted instance f , a deprecated read call to f.fitted_list can be replaced by f.get_fitted_params()['forecasters'] or f.forecasters_ .
[ENH] html display of estimators - tag and consistent handling of via _steps_attr in HeterogenousMetaEstimator descendants ( #7233 ) @mateuszkasprowicz
[ENH] sync dependency checkers with scikit-base ( #7529 ) @fkiraly
[ENH] base design for detection metrics ( #7515 ) @fkiraly
[ENH] refactor InstanceSplitter to a utility apply_split to apply an sklearn split to a time series collection ( #7330 ) @ksharma6
[ENH] Detection metrics - coercion to detection type in boilerplate ( #7546 ) @fkiraly
[ENH] reduce requested metadata in forecasting metric input check ( #7514 ) @fkiraly
[ENH] Expose evaluate params to add_task of forecasting benchmark ( #7574 ) @benHeid
[ENH] Implement efficient _evaluate_by_index for MedianSquaredError ( #7615 ) @satvshr
[ENH] RAND Index metric for time series segmentation ( #7533 ) @gavinkatz001
[ENH] forecasting evaluate : new return_model parameter to return fitted estimator states ( #7637 ) @marrov
[ENH] windowed F1-score for detection ( #7628 ) @gavinkatz001
[ENH] Implement efficient _evaluate_by_index for MdAPE ( #7606 ) @HarshvirSandhu
[ENH] Forecasting metrics dispatch of call to evaluate vs evaluate_by_index ( #7608 ) @benHeid
[ENH] Dataset object interface ( #7398 ) @felipeangelimvieira
[ENH] Remove gluonts dependency in gluonts_ListDataset_panel mtype ( #7558 ) @PranavBhatP
[ENH] data types: early check for python type and host module to improve performance and dependency isolation ( #7736 ) @fkiraly
[ENH] refactor Alignment and Proba datatypes to scikit-base records ( #7739 ) @fkiraly
[ENH] deprecate fitted_list attribute of HierarchyEnsembleForecaster and ( #7423 ) @sanskarmodi8
[ENH] refactor forecasting metric tests prep - object_type tag metric_forecasting ( #7516 ) @fkiraly
[ENH] Add SplineTrendForecaster ( #7487 , #7502 ) @Dehelaan , @jgyasu
[ENH] Interface to skforecast.recursive.ForecasterRecursive ( #7554 ) @yarnabrina
[ENH] naive thresholding detector ( #7576 ) @fkiraly
[ENH] Forecaster imputer for RecursiveReductionForecaster ( #7535 ) @lenaklosik
[ENH] TinyTimeMixer Validation split efficiency fix ( #7647 ) @RHYTHM2405
[ENH] allow arbitrary imputation transformers in 2nd generation reducers ( #7646 ) @fkiraly
[ENH] MAPA forecaster ( #7620 ) @phoeenniixx
[ENH] Set context_len and horizon_len per default in TimesFMForecaster ( #7597 ) @tanvincible
[ENH] Interface to statmodels ar_select_order ( #7693 ) @satvshr
[ENH] change all_estimators retrieval to be entirely tag based ( #7555 ) @fkiraly
[ENH] detection module rework - populate author and maintainer tags ( #7510 ) @fkiraly
[ENH] Fix BaseDetector for segmentation and HMM estimators ( #7480 ) @y-mx , @fkiraly
[ENH] refactor checks for detector outputs into separate module ( #7542 ) @fkiraly
[ENH] detection base class input checks and conversions ( #7577 ) @fkiraly
[ENH] ClaSPSegmentation fixes for new detection interface ( #7585 ) @fkiraly
[ENH] PyODDetector fixes for new detector interface ( #7584 ) @fkiraly
[ENH] window based time series segmentation via clustering ( #7612 ) @Ankit-1204
[ENH] BinarySegmentation fixes for new detection interface ( #7504 ) @Alex-JG3
[ENH] GreedyGaussianSegmentation fixes for new detection interface ( #7472 ) @Spinachboul
[ENH] add sklearn compliance to RotationForest ( #7638 ) @PranavBhatP
[ENH] improvements to clusterer base class and tests ( #7665 ) @fkiraly
[ENH] K-visibility clustering algorithm ( #7592 ) @seigpe
[ENH] clusterers: improved checking for X in fit and predict in capability:out_of_sample = False case, minor improvements to predict_proba default ( #7593 ) @fkiraly
[ENH] Add multiple test parameter sets for YtoX transformer ( #7469 ) @tanvincible
[ENH] create rbf_forecaster.py using RBFTransformer Neural networks ( #7334 ) @phoeenniixx
[ENH] Refactor test class registry to type records ( #7525 ) @fkiraly
[ENH] migrate tests for point forecasting metrics to test class ( #7532 ) @fkiraly
[ENH] clusterer test suite ( #7589 ) @fkiraly
[DOC] Add contributor Gavin Katz to .all-contributorsrc file ( #7513 ) @gavinkatz001
[DOC] Added Note in docstring of InceptionTimeClassifier for fixing #7453 ( #7508 ) @skinan
[DOC] Fixing some typos in transformation.rst ( #7520 ) @phoeenniixx
[DOC] docstring usage examples for AlignerDTW and AlignerDTWfromDist ( #7381 ) @Adarsh2345
[DOC] API reference documentation for SeriesXarray mtype ( #7494 ) @SABARNO-PRAMANICK
[DOC] add proper credits to hfawaz and dl-4-tsc in various time series classifiers ( #7518 ) @fkiraly
[DOC] API specification docstrings for TablePdDataFrame and TablePdSeries mtypes ( #7540 ) @VjayRam
[DOC] API reference for mtype PanelGluontsList ( #7539 ) @b9junkers
[DOC] document new visual_block_kind tag ( #7524 ) @fkiraly
[DOC] API reference for TableNp1D mtype ( #7553 ) @RUPESH-KUMAR01
[DOC] Add missing docstring for functions in /transformations/panel/catch22.py ( #7527 ) @PranavBhatP
[DOC] Added docstrings for TableNp2d and TableListOfDict classes ( #7563 ) @VjayRam
[DOC] minor formatting improvement in hf_transformers_forecaster ( #7659 ) @fkiraly
[DOC] Update the docstring of sameloc splitter with a proper mathematical description ( #7550 ) @keitaVigano
[DOC] AutoREG docstring includes non-existent arguments #7653 ( #7654 ) @cheachu
[DOC] minor improvements in data loader docs ( #7633 ) @fkiraly
[DOC] Fixed TablePolarsEager rendering issues related to recent merged PR #7666 ( #7699 ) @fnhirwa
[DOC] improved formatting and clarity in release manager guide ( #7705 ) @fkiraly
[DOC] improvements to git workflow guide, note about avoiding rebase ( #7670 ) @fkiraly
[DOC] API reference for detection metrics ( #7722 ) @fkiraly
[DOC] improved docstring for grouped/clustering forecaster compositors ( #7709 ) @fkiraly
[DOC] add newer patterns to handle multivariate data in tutorial ( #7362 ) @PranavBhatP
[DOC] update GitHub name of NoaWegerhoff ( #7682 ) @fkiraly
[DOC] Fix typo in example code in documentation requiring double brackets to slice dataframe. DeepAR ( #7641 ) @gbilleyPeco
[DOC] Fix typo in Add Estimators Documentation ( #7519 ) @jgyasu
[DOC] Fixes #7623 Invalid Link ( #7625 ) @jgyasu
[DOC] Added Description for TablePolarsEager ( #7666 ) @RHYTHM2405
[DOC] Fix invalid link to elections repository in sktime’s website ( #7631 ) @tanvincible
[DOC] Fix typo in example code in documentation requiring double brackets to slice dataframe. ( #7639 ) @gbilleyPeco
[DOC] update ttm docstring to inform about zero-shot/fine-tuning ( #7280 ) @geetu040
[MNT] Bump statsforecast to latest version ( #7573 ) @yarnabrina
[MNT] tsfresh estimators - temporary bound on scipy ( #7624 ) @fkiraly
[MNT] temporary skip of test_st_on_unit_test ( #7726 ) @fkiraly
[MNT] fix: add aarch64 installation constraint for temporian ( #7692 ) @abhishek-iitmadras
[MNT] Dependabot: Update pykan requirement from <0.2.7,>=0.2.1 to >=0.2.1,<0.2.9 ( #7396 ) @dependabot[bot]
[MNT] Dependabot: Update dask requirement from <2024.8.1 to <2025.1.1 ( #7481 , #7545 , #7658 ) @dependabot[bot]
[MNT] Dependabot: Update skpro requirement from <2.9.0,>=2 to >=2,<2.10.0 ( #7704 ) @dependabot[bot]
[MNT] Dependabot: Update numba requirement from <0.61 to <0.62 ( #7677 ) @dependabot[bot]
[MNT] Dependabot: Update optuna requirement from <4.2 to <4.3 ( #7676 ) @dependabot[bot]
[BUG] Accepting prereleases as valid python version ( #7544 ) @Abelarm
[BUG] fix get_fitted_params access to wrapped estimators in _HeterogenousEnsembleForecaster descendants ( #7522 ) @fkiraly
[BUG] fix use of Imputer in RecursiveReductionForecaster ( #7706 ) @Salzemann
[BUG] AttentionLSTM - Remove activity regularizers ( #7582 ) @marcosfelt
[BUG] fixing spurious retrieval of fallback Normal distribution ( #7720 ) @fkiraly
[BUG] fix all_estimators lookup in case a tag is used that is not scitype specific ( #7679 ) @fkiraly
[BUG] Ensured that all _fit etc have lower case y ( #7673 ) @KrishBakshi
[BUG] fix expected outputs tests of ShapeletTransformClassifier post shapelet transform fix #7499 ( #7734 ) @fkiraly
[BUG] fix bugs in MACNNClassifier , MACNNRegressor ( #7651 ) @KrishBakshi
[BUG] Fix inconsistency in _online_shapelet_distance std computation ( #7499 ) @fnhirwa
[BUG] TSFreshFeatureExtractor and TSFreshRelevantFeatureExtractor : fix distributor not getting passed through to extract_features ( #7541 ) @marcosfelt
[BUG] in transformers, do not pass y to inner in case of scitype mismatch ( #7733 ) @fkiraly
@Abelarm , @abhishek-iitmadras , @Adarsh2345 , @Alex-JG3 , @Ankit-1204 , @b9junkers , @benHeid , @cheachu , @Dehelaan , @felipeangelimvieira , @fkiraly , @fnhirwa , @gavinkatz001 , @gbilleyPeco , @geetu040 , @HarshvirSandhu , @jgyasu , @keitaVigano , @KrishBakshi , @ksharma6 , @lenaklosik , @marcosfelt , @marrov , @mateuszkasprowicz , @phoeenniixx , @PranavBhatP , @RHYTHM2405 , @RUPESH-KUMAR01 , @SABARNO-PRAMANICK , @Salzemann , @sanskarmodi8 , @satvshr , @seigpe , @skinan , @Spinachboul , @tanvincible , @VjayRam , @y-mx , @yarnabrina
Forecasting metrics __call__ can be passed by_index to return per-index metric evaluate_by_index (7608) benHeid
MAPA forecaster (7620) phoeenniixx
Interface to skforecast.recursive.ForecasterRecursive (7554) yarnabrina
Native reducers (2nd generation) now support arbitrary imputation strategies (7535, 7646) lenaklosik
New detection metrics: RAND Index, windowed F1-score (7533, 7628) gavinkatz001
forecasting evaluate now allows returning fitted estimators via return_model parameter (7637) marrov
Tags and object API for Datasets (7398) felipeangelimvieira
K-visibility clustering algorithm (7592) seigpe
Interface to statmodels ar_select_order for lag order estimation (7693) satvshr
skpro (probability distributions soft dependency) bounds have been updated to >=2,<2.10.0
numba (computation soft dependency) bounds have been updated to <0.62
optuna (hyperparameter optimization soft dependency) bounds have been updated to <4.3
pykan (forecasting soft dependency) bounds have been updated to >=0.2.1,<0.2.9
dask (data container and parallelization back-end soft dependency) bounds have been updated to <2025.1.1
a unified API has been introduced for data sets and data loaders,
following the scikit-base pattern. A unified load method is now available for all data loaders, with further unified API points such as keys and tags for retrieval. For further details, see datasets_ref reference.
All Forecasting metrics can now be constructed with a by_index=True argument.
This will result in a direct call returning the metric per time index. This is equivalent to a call of the evaluate_by_index method. by_index=False is the default, and dispatches, as currently, to the evaluate method of the metric.
In HierarchyEnsembleForecaster, the fitted_list attribute is deprecated. To access the fitted estimators, users should use instead either the get_fitted_params method, or the attribute forecasters_. Given a fitted instance f, a deprecated read call to f.fitted_list can be replaced by f.get_fitted_params()['forecasters'] or f.forecasters_.
[ENH] html display of estimators - tag and consistent handling of via _steps_attr in HeterogenousMetaEstimator descendants (7233) mateuszkasprowicz
[ENH] sync dependency checkers with scikit-base (7529) fkiraly
[ENH] base design for detection metrics (7515) fkiraly
[ENH] refactor InstanceSplitter to a utility apply_split to apply an sklearn split to a time series collection (7330) ksharma6
[ENH] Detection metrics - coercion to detection type in boilerplate (7546) fkiraly
[ENH] reduce requested metadata in forecasting metric input check (7514) fkiraly
[ENH] Expose evaluate params to add_task of forecasting benchmark (7574) benHeid
[ENH] Implement efficient _evaluate_by_index for MedianSquaredError (7615) satvshr
[ENH] RAND Index metric for time series segmentation (7533) gavinkatz001
[ENH] forecasting evaluate: new return_model parameter to return fitted estimator states (7637) marrov
[ENH] windowed F1-score for detection (7628) gavinkatz001
[ENH] Implement efficient _evaluate_by_index for MdAPE (7606) HarshvirSandhu
[ENH] Forecasting metrics dispatch of __call__ to evaluate vs evaluate_by_index (7608) benHeid
[ENH] Dataset object interface (7398) felipeangelimvieira
[ENH] Remove gluonts dependency in gluonts_ListDataset_panel mtype (7558) PranavBhatP
[ENH] data types: early check for python type and host module to improve performance and dependency isolation (7736) fkiraly
[ENH] refactor Alignment and Proba datatypes to scikit-base records (7739) fkiraly
[ENH] deprecate fitted_list attribute of HierarchyEnsembleForecaster and (7423) sanskarmodi8
[ENH] refactor forecasting metric tests prep - object_type tag metric_forecasting (7516) fkiraly
[ENH] Add SplineTrendForecaster (7487, 7502) Dehelaan, jgyasu
[ENH] Interface to skforecast.recursive.ForecasterRecursive (7554) yarnabrina
[ENH] naive thresholding detector (7576) fkiraly
[ENH] Forecaster imputer for RecursiveReductionForecaster (7535) lenaklosik
[ENH] TinyTimeMixer Validation split efficiency fix (7647) RHYTHM2405
[ENH] allow arbitrary imputation transformers in 2nd generation reducers (7646) fkiraly
[ENH] MAPA forecaster (7620) phoeenniixx
[ENH] Set context_len and horizon_len per default in TimesFMForecaster (7597) tanvincible
[ENH] Interface to statmodels ar_select_order (7693) satvshr
[ENH] change all_estimators retrieval to be entirely tag based (7555) fkiraly
[ENH] detection module rework - populate author and maintainer tags (7510) fkiraly
[ENH] Fix BaseDetector for segmentation and HMM estimators (7480) y-mx, fkiraly
[ENH] refactor checks for detector outputs into separate module (7542) fkiraly
[ENH] detection base class input checks and conversions (7577) fkiraly
[ENH] ClaSPSegmentation fixes for new detection interface (7585) fkiraly
[ENH] PyODDetector fixes for new detector interface (7584) fkiraly
[ENH] window based time series segmentation via clustering (7612) Ankit-1204
[ENH] BinarySegmentation fixes for new detection interface (7504) Alex-JG3
[ENH] GreedyGaussianSegmentation fixes for new detection interface (7472) Spinachboul
[ENH] add sklearn compliance to RotationForest (7638) PranavBhatP
[ENH] improvements to clusterer base class and tests (7665) fkiraly
[ENH] K-visibility clustering algorithm (7592) seigpe
[ENH] clusterers: improved checking for X in fit and predict in capability:out_of_sample = False case, minor improvements to predict_proba default (7593) fkiraly
[ENH] Add multiple test parameter sets for YtoX transformer (7469) tanvincible
[ENH] create rbf_forecaster.py using RBFTransformer Neural networks (7334) phoeenniixx
[ENH] Refactor test class registry to type records (7525) fkiraly
[ENH] migrate tests for point forecasting metrics to test class (7532) fkiraly
[ENH] clusterer test suite (7589) fkiraly
[DOC] Add contributor Gavin Katz to .all-contributorsrc file (7513) gavinkatz001
[DOC] Added Note in docstring of InceptionTimeClassifier for fixing #7453 (7508) skinan
[DOC] Fixing some typos in transformation.rst (7520) phoeenniixx
[DOC] docstring usage examples for AlignerDTW and AlignerDTWfromDist (7381) Adarsh2345
[DOC] API reference documentation for SeriesXarray mtype (7494) SABARNO-PRAMANICK
[DOC] add proper credits to hfawaz and dl-4-tsc in various time series classifiers (7518) fkiraly
[DOC] API specification docstrings for TablePdDataFrame and TablePdSeries mtypes (7540) VjayRam
[DOC] API reference for mtype PanelGluontsList (7539) b9junkers
[DOC] document new visual_block_kind tag (7524) fkiraly
[DOC] API reference for TableNp1D mtype (7553) RUPESH-KUMAR01
[DOC] Add missing docstring for functions in /transformations/panel/catch22.py (7527) PranavBhatP
[DOC] Added docstrings for TableNp2d and TableListOfDict classes (7563) VjayRam
[DOC] minor formatting improvement in hf_transformers_forecaster (7659) fkiraly
[DOC] Update the docstring of sameloc splitter with a proper mathematical description (7550) keitaVigano
[DOC] AutoREG docstring includes non-existent arguments #7653 (7654) cheachu
[DOC] minor improvements in data loader docs (7633) fkiraly
[DOC] Fixed TablePolarsEager rendering issues related to recent merged PR #7666 (7699) fnhirwa
[DOC] improved formatting and clarity in release manager guide (7705) fkiraly
[DOC] improvements to git workflow guide, note about avoiding rebase (7670) fkiraly
[DOC] API reference for detection metrics (7722) fkiraly
[DOC] improved docstring for grouped/clustering forecaster compositors (7709) fkiraly
[DOC] add newer patterns to handle multivariate data in tutorial (7362) PranavBhatP
[DOC] update GitHub name of NoaWegerhoff (7682) fkiraly
[DOC] Fix typo in example code in documentation requiring double brackets to slice dataframe. DeepAR (7641) gbilleyPeco
[DOC] Fix typo in Add Estimators Documentation (7519) jgyasu
[DOC] Fixes #7623 Invalid Link (7625) jgyasu
[DOC] Added Description for TablePolarsEager (7666) RHYTHM2405
[DOC] Fix invalid link to elections repository in sktime's website (7631) tanvincible
[DOC] Fix typo in example code in documentation requiring double brackets to slice dataframe. (7639) gbilleyPeco
[DOC] update ttm docstring to inform about zero-shot/fine-tuning (7280) geetu040
[MNT] Bump statsforecast to latest version (7573) yarnabrina
[MNT] tsfresh estimators - temporary bound on scipy (7624) fkiraly
[MNT] temporary skip of test_st_on_unit_test (7726) fkiraly
[MNT] fix: add aarch64 installation constraint for temporian (7692) abhishek-iitmadras
[MNT] [Dependabot](deps): Update pykan requirement from <0.2.7,>=0.2.1 to >=0.2.1,<0.2.9 (7396) dependabot[bot]
[MNT] [Dependabot](deps): Update dask requirement from <2024.8.1 to <2025.1.1 (7481, 7545, 7658) dependabot[bot]
[MNT] [Dependabot](deps): Update skpro requirement from <2.9.0,>=2 to >=2,<2.10.0 (7704) dependabot[bot]
[MNT] [Dependabot](deps): Update numba requirement from <0.61 to <0.62 (7677) dependabot[bot]
[MNT] [Dependabot](deps): Update optuna requirement from <4.2 to <4.3 (7676) dependabot[bot]
[BUG] Accepting prereleases as valid python version (7544) Abelarm
[BUG] fix get_fitted_params access to wrapped estimators in _HeterogenousEnsembleForecaster descendants (7522) fkiraly
[BUG] fix use of Imputer in RecursiveReductionForecaster (7706) Salzemann
[BUG] AttentionLSTM - Remove activity regularizers (7582) marcosfelt
[BUG] fixing spurious retrieval of fallback Normal distribution (7720) fkiraly
[BUG] fix all_estimators lookup in case a tag is used that is not scitype specific (7679) fkiraly
[BUG] Ensured that all _fit etc have lower case y (7673) KrishBakshi
[BUG] fix expected outputs tests of ShapeletTransformClassifier post shapelet transform fix #7499 (7734) fkiraly
[BUG] fix bugs in MACNNClassifier, MACNNRegressor (7651) KrishBakshi
[BUG] Fix inconsistency in _online_shapelet_distance std computation (7499) fnhirwa
[BUG] TSFreshFeatureExtractor and TSFreshRelevantFeatureExtractor: fix distributor not getting passed through to extract_features (7541) marcosfelt
[BUG] in transformers, do not pass y to inner in case of scitype mismatch (7733) fkiraly
Abelarm, abhishek-iitmadras, Adarsh2345, Alex-JG3, Ankit-1204, b9junkers, benHeid, cheachu, Dehelaan, felipeangelimvieira, fkiraly, fnhirwa, gavinkatz001, gbilleyPeco, geetu040, HarshvirSandhu, jgyasu, keitaVigano, KrishBakshi, ksharma6, lenaklosik, marcosfelt, marrov, mateuszkasprowicz, phoeenniixx, PranavBhatP, RHYTHM2405, RUPESH-KUMAR01, SABARNO-PRAMANICK, Salzemann, sanskarmodi8, satvshr, seigpe, skinan, Spinachboul, tanvincible, VjayRam, y-mx, yarnabrina
Maintenance release with scheduled deprecations and change actions.
Maintenance release with scheduled deprecations and change actions.
For last larger feature update, see 0.34.1.
Please see our changelog for a description of all changes.
@fkiraly, @fnhirwa, @yarnabrina
Full Changelog: https://github.com/sktime/sktime/compare/v0.34.1...v0.35.0
Maintenance release with scheduled deprecations and change actions.
For the last non-maintenance content update, see 0.34.1.
scikit-base bounds have been updated to >=0.6.1,<0.13.0
skpro (probability distributions soft dependency) bounds have been updated to >=2,<2.9.0
From sktime 0.38.0 , forecasters’ predict_proba will require skpro to be present in the python environment, for distribution objects to represent distributional forecasts. Since sktime 0.35.0 , an error is raised upon call of forecaster predict_proba if skpro is not present in the environment. Users of forecasters’ predict_proba should ensure that skpro is installed in the environment.
The probability distributions module sktime.proba deprecated and will be fully replaced by skpro in sktime 0.38.0 . Until sktime 0.38.0 , imports from sktime.proba will continue working, defaulting to sktime.proba if skpro is not present, otherwise redirecting imports to skpro objects. Since sktime 0.35.0 , an error is raised if skpro is not present in the environment, otherwise imports are redirected to skpro . Direct or indirect users of sktime.proba should ensure skpro is installed in the environment. Direct users of the sktime.proba module should, in addition, replace any imports from sktime.proba with imports from skpro.distributions .
in neuralforecast facing estimator interfaces, the default for the broadcasting parameter has been consistently changed to False .
[MNT] Dependabot: Update scikit-base requirement from <0.12.0,>=0.6.1 to >=0.6.1,<0.13.0 ( #7392 ) @dependabot[bot]
[MNT] Dependabot: Update skpro requirement from <2.8.0,>=2 to >=2,<2.9.0 ( #7406 ) @dependabot[bot]
[MNT] bound peft<0.14.0 ( #7495 ) @fkiraly
[MNT] 0.35.0 deprecations and change actions ( #7485 ) @fkiraly
[MNT] remove top level import of transformers ( #7491 ) @yarnabrina
[MNT] add upper bound to skforecast autoreg adapter ( #7488 ) @yarnabrina
[DOC] fix failing cells in detection notebook after skchange update ( #7483 ) @fkiraly
[BUG] Fix codec error on traditional Chinese Windows systems ( #7498 ) @fnhirwa
@fkiraly , @fnhirwa , @yarnabrina
Maintenance release with scheduled deprecations and change actions.
For the last non-maintenance content update, see 0.34.1.
scikit-base bounds have been updated to >=0.6.1,<0.13.0
skpro (probability distributions soft dependency) bounds have been updated to >=2,<2.9.0
From sktime 0.38.0, forecasters' predict_proba will require skpro to be present in the python environment, for distribution objects to represent distributional forecasts. Since sktime 0.35.0, an error is raised upon call of forecaster predict_proba if skpro is not present in the environment. Users of forecasters' predict_proba should ensure that skpro is installed in the environment.
The probability distributions module sktime.proba deprecated and will be fully replaced by skpro in sktime 0.38.0. Until sktime 0.38.0, imports from sktime.proba will continue working, defaulting to sktime.proba if skpro is not present, otherwise redirecting imports to skpro objects. Since sktime 0.35.0, an error is raised if skpro is not present in the environment, otherwise imports are redirected to skpro. Direct or indirect users of sktime.proba should ensure skpro is installed in the environment. Direct users of the sktime.proba module should, in addition, replace any imports from sktime.proba with imports from skpro.distributions.
in neuralforecast facing estimator interfaces, the default for the broadcasting parameter has been consistently changed to False.
[MNT] [Dependabot](deps): Update scikit-base requirement from <0.12.0,>=0.6.1 to >=0.6.1,<0.13.0 (7392) dependabot[bot]
[MNT] [Dependabot](deps): Update skpro requirement from <2.8.0,>=2 to >=2,<2.9.0 (7406) dependabot[bot]
[MNT] bound peft<0.14.0 (7495) fkiraly
[MNT] 0.35.0 deprecations and change actions (7485) fkiraly
[MNT] remove top level import of transformers (7491) yarnabrina
[MNT] add upper bound to skforecast autoreg adapter (7488) yarnabrina
[DOC] fix failing cells in detection notebook after skchange update (7483) fkiraly
[BUG] Fix codec error on traditional Chinese Windows systems (7498) fnhirwa
fkiraly, fnhirwa, yarnabrina
Please see our changelog for a description of all changes.
Feature release.
Please see our changelog for a description of all changes.
@Akhsuna07, @Alex-JG3, @alyssadsouza, @Dehelaan, @ericjb, @fkiraly, @gavinkatz001, @geetu040, @hudsonhoch, @jgyfutub, @julian-fong, @jusssch, @keitaVigano, @liya-zhu, @manolotis, @MarkusSagen, @medha-14, @mjste, @pranavvp16, @sanskarmodi8, @ShivamJ07, @Sohaib-Ahmed21, @SSROCKS30, @tajir0, @talat-khattatov, @vagechirkov, @VectorNd, @yarnabrina
Full Changelog: https://github.com/sktime/sktime/compare/v0.34.0...v0.34.1
Full rework of detectors API, module is now maturing. @tveten , @alex-jg3 , @fkiraly , @alyssadsouza , @ShivamJ07 , @RobotPsychologist
in forecasting metrics, allow a callable to be passed as sample_weight for dynamic weight generation ( #7288 ) @MarkusSagen
Interfaces to new neuralforecast estimators ( #7434 ) @yarnabrina
Box-Cox bias adjustment for forecasters ( #7268 ) @sanskarmodi8 , @talat-khattatov
SCINet forecaster ( #7400 ) @Sohaib-Ahmed21
Spatio-Temporal DBSCAN clusterer ( #7353 ) @vagechirkov , @vectornd
Signature moments transformer ( #7320 ) @VectorNd
dask (data container and parallelization back-end soft dependency) bounds have been updated to <2024.11.3
Forecasting metrics now allow a callable to be passed as sample_weight for dynamic weight generation. The precise contract the callables need to satisfy is documented in the docstrings.
The core API for time series detectors has changed, and has been homogenized with the API of skchange .
The changes are in parts breaking, and require adjustments in the user code. The changes were necessary to ensure a consistent API for time series detectors.
BaseDetector.predict now returns a pd.DataFrame . This DataFrame always has a RangeIndex , with individual rows corresponding to individual events. The DataFrame has at least a column "ilocs" , which contains the integer indices of the detected events, in reference to the argument X passed to predict . The format, for an entry, can be a single integer or a left-closed pd.Interval of integers, indicating either a single point or a segment of points.
an additional column "labels" may be present in cases of labelled events.
BaseDetector.predict_points and predict_segments now also return pd.DataFrame , following the same format as predict , with the difference that predict_points coerces detected events to points, taking start and end points of segments as separate points. predict_segments coerces detected events to segments, possibly single-index segments.
BaseDetector.transform now always returns a pd.DataFrame with, which has at least a column "transform" containing the transformed series.
[ENH] refactor _check_estimator_types to use record class interface ( #7395 ) @fkiraly
[ENH] parity of _HeterogenousMetaEstimator._check_estimator with skbase and fixes in EnsembleForecaster.init ( #7429 ) @fkiraly
[ENH] MSRE-percent forecasting metric ( #7394 ) @jusssch , @fkiraly
[ENH] add RecursiveReductionForecaster to module exports, tests, and documentation ( #4806 ) @fkiraly
[ENH] EnsembleForecaster option to specify multiple copies of same forecaster easily ( #7424 ) @ericjb
[ENH] Add new arguments to neuralforecast rnn and lstm ( #7422 ) @yarnabrina
[ENH] Box-Cox bias adjustment for forecasters ( #7268 ) @sanskarmodi8 , @talat-khattatov
[ENH] in forecasting metrics, allow a callable to be passed as sample_weight for dynamic weight generation ( #7288 ) @MarkusSagen
[ENH] SCINet forecaster ( #7400 ) @Sohaib-Ahmed21
[ENH] Reduce memory footprint of statsmodels adapter - save necessary _y info to remove need for _y ( #7417 ) @hudsonhoch
[ENH] Interface new neuralforecast estimators ( #7434 ) @yarnabrina
[ENH] Test parameter set for STRAY ( #7420 ) @tajir0
[ENH] add _transform_score template and remove dead fmt related code from detectors ( #7425 ) @ShivamJ07
[ENH] homogenization of sktime and skchange detection API - base class predict return type ( #7433 ) @alyssadsouza
[ENH] rename private make_annotation_problem to make_detection_problem ( #7436 ) @fkiraly
[ENH] minor improvements to BaseDetector ( #7435 ) @fkiraly
[ENH] move detector non-suite tests to detection folder ( #7445 ) @fkiraly
[ENH] homogenization of sktime and skchange detection API - part 1 - base class and y argument ( #7342 ) @fkiraly
[ENH] homogenization of sktime and skchange detection API - part 2 - base class transform return type ( #7432 ) @alyssadsouza
[ENH] move detector classes to detection module ( #7448 ) @fkiraly
[ENH] conditional testing and imports for estimator specific tests in detection module ( #7446 ) @fkiraly
[ENH] add detector specific tags to the tag registry and tags API reference ( #7443 ) @liya-zhu
[ENH] detector dummies ( #7440 ) @fkiraly
[ENH] Changed scitype name "series-annotator" to "detector" ( #7361 ) @jgyfutub
[ENH] homogenization of sktime and skchange detection API - base class predict_scores return type ( #7460 ) @alyssadsouza
[ENH] homogenization of sktime and skchange detection API - base class predict_points return type ( #7459 ) @alyssadsouza
[ENH] skchange rename of Capa to CAPA ( #7457 ) @fkiraly
[ENH] move detector tests to detection module ( #7455 , #7464 ) @fkiraly
[ENH] renaming of detector suite tests ( #7465 ) @fkiraly
[ENH] fix API incompatibility of SubLOF detector ( #7468 ) @fkiraly
[ENH] add skchange change point detectors as placeholder records ( #7458 ) @ShivamJ07 , @fkiraly
[ENH] add placeholder records for skchange segment anomaly detectors ( #7470 ) @fkiraly
[ENH] update DetectorPipeline and DetectorAsTransformer inner calls and docstrings ( #7475 ) @fkiraly
[ENH] homogenization of sktime and skchange detection API - predict , predict_points , predict_segments ( #7476 ) @fkiraly
[ENH] second test parameter to Arsenal classifier ( #7335 ) @jusssch
[ENH] Spatio-Temporal DBSCAN clusterer ( #7353 ) @vagechirkov , @vectornd
[ENH] Test parameter sets for ComposableTimeSeriesForestRegressor ( #7431 ) @Dehelaan
[ENH] Signature Moments transformer ( #7320 ) @VectorNd
[ENH] added test parameter set for ColumnTransformer ( #7442 ) @sanskarmodi8
[ENH] second test parameter set in _multirocket_multivariate ( #7344 ) @medha-14
[ENH] detection module suite tests coverage extension for methods ( #6958 ) @Alex-JG3
[DOC] fix typos in changelog, better explanation of scoring parameters ( #7324 ) @fkiraly
[DOC] in docstring, rename Example to Examples sections ( #7341 ) @fkiraly
[DOC] add docstring example for DropNA ( #7327 ) @SSROCKS30
[DOC] Fixed sktime logo positioning in the README ( #7322 ) @Akhsuna07
[DOC] Add uni2ts in libs/README.md ( #7348 ) @pranavvp16
[DOC] Small typo on explanation of regression ( #7366 ) @manolotis
[DOC] Add Binary Segmentation Estimator to API reference ( #7379 ) @Dehelaan
[DOC] Add timesfm in libs/README.md ( #7358 ) @geetu040
[DOC] categories for clustering algorithms in API reference ( #7369 ) @fkiraly
[DOC] API reference for data representations ( #7231 ) @fkiraly
[DOC] Improving the installation instructions to improve clarity ( #7339 ) @julian-fong
[DOC] fixes to the mtype API reference ( #7393 ) @fkiraly
[DOC] fix duplications of the word “correspond” ( #7403 ) @fkiraly
[DOC] Add import location to HolidayFeatures docs ( #7401 ) @mjste
[DOC] add rubric for time series detectors to estimator overview ( #7389 ) @gavinkatz001
[DOC] provide proper mathematical description in docstring of sliding window splitter. ( #7376 ) @keitaVigano
[DOC] add detector tags to estimator overview, fix API reference ( #7473 ) @fkiraly
[DOC] improve formatting of forecasting metrics ( #7474 ) @fkiraly
[MNT] allow optuna<4.2 and test 3.13 support ( #7384 ) @fkiraly
[MNT] detection module - move CI from annotation module ( #7456 ) @fkiraly
[MNT] updating versions of pre-commit hooks ( #7461 ) @yarnabrina
[MNT] Dependabot: Update dask requirement from <2024.8.1 to <2024.11.3 ( #7387 ) @dependabot[bot]
[MNT] Dependabot: Update pytest-randomly requirement from <3.16,>=3.15 to >=3.15,<3.17 ( #7338 ) @dependabot[bot]
[BUG] rectify ignores-exogeneous-X tag value in Croston ( #7404 ) @yarnabrina
[BUG] fix BaseDetector methods for case of no points detected ( #7439 ) @fkiraly
[BUG] ensure BaseDetector.transform uses iloc indexing ( #7444 ) @fkiraly
[BUG] fix BaseDetector.change_points_to_segments ( #7452 ) @fkiraly
[BUG] fix input shape in CNTCClassifier and CNTCRegressor ( #7269 ) @sanskarmodi8
[BUG] fix input shape in CNTCClassifier ( #7269 ) @sanskarmodi8
[BUG] fix input shape in CNTCRegressor ( #7269 ) @sanskarmodi8
[BUG] fix check_estimator test in case of failure ( #7430 ) @fkiraly
@Akhsuna07 , @Alex-JG3 , @alyssadsouza , @Dehelaan , @ericjb , @fkiraly , @gavinkatz001 , @geetu040 , @hudsonhoch , @jgyfutub , @julian-fong , @jusssch , @keitaVigano , @liya-zhu , @manolotis , @MarkusSagen , @medha-14 , @mjste , @pranavvp16 , @sanskarmodi8 , @ShivamJ07 , @Sohaib-Ahmed21 , @SSROCKS30 , @tajir0 , @talat-khattatov , @vagechirkov , @VectorNd , @yarnabrina
Full rework of detectors API, module is now maturing. tveten, alex-jg3, fkiraly, alyssadsouza, ShivamJ07, RobotPsychologist
in forecasting metrics, allow a callable to be passed as sample_weight for dynamic weight generation (7288) MarkusSagen
Interfaces to new neuralforecast estimators (7434) yarnabrina
Box-Cox bias adjustment for forecasters (7268) sanskarmodi8, talat-khattatov
SCINet forecaster (7400) Sohaib-Ahmed21
Spatio-Temporal DBSCAN clusterer (7353) vagechirkov, vectornd
Signature moments transformer (7320) VectorNd
dask (data container and parallelization back-end soft dependency) bounds have been updated to <2024.11.3
Forecasting metrics now allow a callable to be passed as sample_weight for dynamic weight generation. The precise contract the callables need to satisfy is documented in the docstrings.
The core API for time series detectors has changed, and has been homogenized with the API of skchange.
The changes are in parts breaking, and require adjustments in the user code. The changes were necessary to ensure a consistent API for time series detectors.
BaseDetector.predict now returns a pd.DataFrame. This DataFrame always has a RangeIndex, with individual rows corresponding to individual events. The DataFrame has at least a column "ilocs", which contains the integer indices of the detected events, in reference to the argument X passed to predict. The format, for an entry, can be a single integer or a left-closed pd.Interval of integers, indicating either a single point or a segment of points.
an additional column "labels" may be present in cases of labelled events.
BaseDetector.predict_points and predict_segments now also return pd.DataFrame, following the same format as predict, with the difference that predict_points coerces detected events to points, taking start and end points of segments as separate points. predict_segments coerces detected events to segments, possibly single-index segments.
BaseDetector.transform now always returns a pd.DataFrame with, which has at least a column "transform" containing the transformed series.
[ENH] refactor _check_estimator_types to use record class interface (7395) fkiraly
[ENH] parity of _HeterogenousMetaEstimator._check_estimator with skbase and fixes in EnsembleForecaster.__init__ (7429) fkiraly
[ENH] MSRE-percent forecasting metric (7394) jusssch, fkiraly
[ENH] add RecursiveReductionForecaster to module exports, tests, and documentation (4806) fkiraly
[ENH] EnsembleForecaster option to specify multiple copies of same forecaster easily (7424) ericjb
[ENH] Add new arguments to neuralforecast rnn and lstm (7422) yarnabrina
[ENH] Box-Cox bias adjustment for forecasters (7268) sanskarmodi8, talat-khattatov
[ENH] in forecasting metrics, allow a callable to be passed as sample_weight for dynamic weight generation (7288) MarkusSagen
[ENH] SCINet forecaster (7400) Sohaib-Ahmed21
[ENH] Reduce memory footprint of statsmodels adapter - save necessary _y info to remove need for _y (7417) hudsonhoch
[ENH] Interface new neuralforecast estimators (7434) yarnabrina
[ENH] Test parameter set for STRAY (7420) tajir0
[ENH] add _transform_score template and remove dead fmt related code from detectors (7425) ShivamJ07
[ENH] homogenization of sktime and skchange detection API - base class predict return type (7433) alyssadsouza
[ENH] rename private make_annotation_problem to make_detection_problem (7436) fkiraly
[ENH] minor improvements to BaseDetector (7435) fkiraly
[ENH] move detector non-suite tests to detection folder (7445) fkiraly
[ENH] homogenization of sktime and skchange detection API - part 1 - base class and y argument (7342) fkiraly
[ENH] homogenization of sktime and skchange detection API - part 2 - base class transform return type (7432) alyssadsouza
[ENH] move detector classes to detection module (7448) fkiraly
[ENH] conditional testing and imports for estimator specific tests in detection module (7446) fkiraly
[ENH] add detector specific tags to the tag registry and tags API reference (7443) liya-zhu
[ENH] detector dummies (7440) fkiraly
[ENH] Changed scitype name "series-annotator" to "detector" (7361) jgyfutub
[ENH] homogenization of sktime and skchange detection API - base class predict_scores return type (7460) alyssadsouza
[ENH] homogenization of sktime and skchange detection API - base class predict_points return type (7459) alyssadsouza
[ENH] skchange rename of Capa to CAPA (7457) fkiraly
[ENH] move detector tests to detection module (7455, 7464) fkiraly
[ENH] renaming of detector suite tests (7465) fkiraly
[ENH] fix API incompatibility of SubLOF detector (7468) fkiraly
[ENH] add skchange change point detectors as placeholder records (7458) ShivamJ07, fkiraly
[ENH] add placeholder records for skchange segment anomaly detectors (7470) fkiraly
[ENH] update DetectorPipeline and DetectorAsTransformer inner calls and docstrings (7475) fkiraly
[ENH] homogenization of sktime and skchange detection API - predict, predict_points, predict_segments (7476) fkiraly
[ENH] second test parameter to Arsenal classifier (7335) jusssch
[ENH] Spatio-Temporal DBSCAN clusterer (7353) vagechirkov, vectornd
[ENH] Test parameter sets for ComposableTimeSeriesForestRegressor (7431) Dehelaan
[ENH] Signature Moments transformer (7320) VectorNd
[ENH] added test parameter set for ColumnTransformer (7442) sanskarmodi8
[ENH] second test parameter set in _multirocket_multivariate (7344) medha-14
[ENH] detection module suite tests coverage extension for methods (6958) Alex-JG3
[DOC] fix typos in changelog, better explanation of scoring parameters (7324) fkiraly
[DOC] in docstring, rename Example to Examples sections (7341) fkiraly
[DOC] add docstring example for DropNA (7327) SSROCKS30
[DOC] Fixed sktime logo positioning in the README (7322) Akhsuna07
[DOC] Add uni2ts in libs/README.md (7348) pranavvp16
[DOC] Small typo on explanation of regression (7366) manolotis
[DOC] Add Binary Segmentation Estimator to API reference (7379) Dehelaan
[DOC] Add timesfm in libs/README.md (7358) geetu040
[DOC] categories for clustering algorithms in API reference (7369) fkiraly
[DOC] API reference for data representations (7231) fkiraly
[DOC] Improving the installation instructions to improve clarity (7339) julian-fong
[DOC] fixes to the mtype API reference (7393) fkiraly
[DOC] fix duplications of the word "correspond" (7403) fkiraly
[DOC] Add import location to HolidayFeatures docs (7401) mjste
[DOC] add rubric for time series detectors to estimator overview (7389) gavinkatz001
[DOC] provide proper mathematical description in docstring of sliding window splitter. (7376) keitaVigano
[DOC] add detector tags to estimator overview, fix API reference (7473) fkiraly
[DOC] improve formatting of forecasting metrics (7474) fkiraly
[MNT] allow optuna<4.2 and test 3.13 support (7384) fkiraly
[MNT] detection module - move CI from annotation module (7456) fkiraly
[MNT] updating versions of pre-commit hooks (7461) yarnabrina
[MNT] [Dependabot](deps): Update dask requirement from <2024.8.1 to <2024.11.3 (7387) dependabot[bot]
[MNT] [Dependabot](deps): Update pytest-randomly requirement from <3.16,>=3.15 to >=3.15,<3.17 (7338) dependabot[bot]
[BUG] rectify ignores-exogeneous-X tag value in Croston (7404) yarnabrina
[BUG] fix BaseDetector methods for case of no points detected (7439) fkiraly
[BUG] ensure BaseDetector.transform uses iloc indexing (7444) fkiraly
[BUG] fix BaseDetector.change_points_to_segments (7452) fkiraly
[BUG] fix input shape in CNTCClassifier and CNTCRegressor (7269) sanskarmodi8
[BUG] fix input shape in CNTCClassifier (7269) sanskarmodi8
[BUG] fix input shape in CNTCRegressor (7269) sanskarmodi8
[BUG] fix check_estimator test in case of failure (7430) fkiraly
Akhsuna07, Alex-JG3, alyssadsouza, Dehelaan, ericjb, fkiraly, gavinkatz001, geetu040, hudsonhoch, jgyfutub, julian-fong, jusssch, keitaVigano, liya-zhu, manolotis, MarkusSagen, medha-14, mjste, pranavvp16, sanskarmodi8, ShivamJ07, Sohaib-Ahmed21, SSROCKS30, tajir0, talat-khattatov, vagechirkov, VectorNd, yarnabrina
scheduled deprecations and change actions.
Maintenance release:
full support for python 3.13
scheduled deprecations and change actions.
For last non-maintenance content updates, see 0.33.2.
numpy bounds have been updated to >=1.21,<2.2
scikit-base bounds have been updated to >=0.6.1,<0.12.0
holidays (transformations soft dependency) bounds have been updated to >=0.29,<0.59
The annotation module will be renamed to detection , starting with 0.34.0. Until a future release not earlier than 0.37.0 or 1.0.0, the module will be available under both names to ensure downwards compatibility of imports.
The annotation soft dependency set will be renamed to detection . The annotation soft dependency set will be available until a future release not earlier than 0.37.0 or 1.0.0, to ensure downwards compatibility of imports.
[MNT] python 3.13 support in pyproject.toml and CI ( #7198 ) @fkiraly
[MNT] disable numpy<2 bound for prophet ( #6740 ) @fkiraly
[MNT] 0.34.0 deprecations and change actions ( #7302 ) @fkiraly
[MNT] prepare rename annotation module to detection - mapped imports ( #7294 ) @fkiraly
[MNT] deduplicate sktime and skbase BaseEstimator ( #7213 ) @fkiraly
[MNT] Dependabot: Update numpy requirement from <2.1,>=1.21 to >=1.21,<2.2 ( #7103 ) @dependabot[bot]
[MNT] Dependabot: Update scikit-base requirement from <0.9.0,>=0.6.1 to >=0.6.1,<0.12.0 ( #7238 ) @dependabot[bot]
[MNT] Dependabot: Update holidays requirement from <0.55,>=0.29 to >=0.29,<0.59 ( #7308 ) @dependabot[bot]
[MNT] Dependabot: Update sphinx-issues requirement from <5.0.0 to <6.0.0 ( #7307 ) @dependabot[bot]
[DOC] replace documentation references to “annotation” with “detection” ( #7299 ) @fkiraly
[DOC] change imports in detection tutorial to detection module ( #7306 ) @fkiraly
Maintenance release:
full support for python 3.13
scheduled deprecations and change actions.
For last non-maintenance content updates, see 0.33.2.
numpy bounds have been updated to >=1.21,<2.2
scikit-base bounds have been updated to >=0.6.1,<0.12.0
holidays (transformations soft dependency) bounds have been updated to >=0.29,<0.59
The annotation module will be renamed to detection, starting with 0.34.0. Until a future release not earlier than 0.37.0 or 1.0.0, the module will be available under both names to ensure downwards compatibility of imports.
The annotation soft dependency set will be renamed to detection. The annotation soft dependency set will be available until a future release not earlier than 0.37.0 or 1.0.0, to ensure downwards compatibility of imports.
[MNT] python 3.13 support in pyproject.toml and CI (7198) fkiraly
[MNT] disable numpy<2 bound for prophet (6740) fkiraly
[MNT] 0.34.0 deprecations and change actions (7302) fkiraly
[MNT] prepare rename annotation module to detection - mapped imports (7294) fkiraly
[MNT] deduplicate sktime and skbase BaseEstimator (7213) fkiraly
[MNT] [Dependabot](deps): Update numpy requirement from <2.1,>=1.21 to >=1.21,<2.2 (7103) dependabot[bot]
[MNT] [Dependabot](deps): Update scikit-base requirement from <0.9.0,>=0.6.1 to >=0.6.1,<0.12.0 (7238) dependabot[bot]
[MNT] [Dependabot](deps): Update holidays requirement from <0.55,>=0.29 to >=0.29,<0.59 (7308) dependabot[bot]
[MNT] [Dependabot](deps): Update sphinx-issues requirement from <5.0.0 to <6.0.0 (7307) dependabot[bot]
[DOC] replace documentation references to "annotation" with "detection" (7299) fkiraly
[DOC] change imports in detection tutorial to detection module (7306) fkiraly
Please see our changelog for a description of all changes.
Feature release.
Please see our changelog for a description of all changes.
@alex-jg3, @Anuragwagh, @benHeid, @Dehelaan, @ericjb, @fkiraly, @fnhirwa, @Garve, @geetu040, @Humorloos, @jan-mue, @julian-fong, @KarlKolibri, @MarkusSagen, @ninedigits, @phoeenniixx, @Prtm2110, @RobotPsychologist, @rigvedmanoj, @SaiRevanth25, @sanskarmodi8, @shivanshsinghal-22, @Smoothengineer, @talat-khattatov, @tianjiqx, @vedantag17, @XinyuWuu, @Z-Fran
Full Changelog: https://github.com/sktime/sktime/compare/v0.33.1...v0.33.2
new forecaster: interface to Chronos (zero-shot) foundation model forecaster ( #7001 ) @Z-Fran , @geetu040 , @benHeid , @rigvedmanoj
new classifiers: GRU-based time series classifiers ( #6952 ) @fnhirwa
new transformer: temporal radial basis function features ( #7261 ) @phoeenniixx
new ExpandingSlidingWindowSplitter , switching from expanding window to sliding window at cutoff point ( #7193 ) @MarkusSagen
pytorch-forecasting and neuralforecast models now provide probabilistic forecasts for global forecasting ( #6628 , #6666 ) @XinyuWuu
PolynomialTrendForecaster now can make probabilistic forecasts ( #6424 ) @ericjb
TimesFMForecaster now allows to select the source package, and zero-shot usage has been memory optimized ( #7204 , #7212 ) @Prtm2110 , @fkiraly
skpro (forecasting soft dependency) bounds have been updated to >=2,<2.8.0
u8darts (forecasting soft dependency) bounds have been updated to >=0.29.0,<0.32.0
dask (data container and parallelization back-end soft dependency) bounds have been updated to <2024.10.1
Performance metrics for probabilistic forecasting, for interval and quantile returns, can now optionally be constructed with an alpha (quantile forecasts) or coverage (interval forecasts) parameter, for example EmpiricalCoverage(coverage=0.7) for “empirical coverage at nominal interval coverage 70 percent”.
These are for use with benchmarking or tuning interfaces, where a metric is provided, but predict_interval or predict_quantiles are not explicitly called.
In such a case, the parameter of the metric will be used by the tuner or benchmark utility to tune with the metric at that quantile alpha or interval coverage .
For example, ForecastingGridSearchCV(fcst, cv, scoring=PinballLoss(alpha=[0.3, 0.7])) , to tune the forecaster fcst scored by the pinball loss at quantiles 0.3 and 0.7.
[ENH] version tag for objects and estimators ( #3629 ) @fkiraly
[ENH] add test that html repr of objects does not crash ( #7151 ) @fkiraly
[ENH] decorators for singleton and multiton oop pattern ( #7203 ) @fkiraly
[ENH] ExpandingSlidingWindowSplitter , switching from expanding window to sliding window at cutoff point ( #7193 ) @MarkusSagen
[ENH] added coverage parameter to all metrics of pred_interval type ( #7278 ) @talat-khattatov
[ENH] refactor datatypes module to scikit-base classes and data records ( #7161 ) @fkiraly
[ENH] Fix polars PerformanceWarning when object is of type polars.LazyFrame ( #7221 ) @shivanshsinghal-22
[ENH] Added method predict_interval to PolynomialTrendForecaster ( #6424 ) @ericjb
[ENH] Quantile Forecast for pytorch-forecasting Models with Global Forecast API ( #6628 ) @XinyuWuu
[ENH] Global Forecast API for NeuralForecast interface ( #6666 ) @XinyuWuu
[ENH] testing global forecasters: reduces number of obs in _make_hierarchical for test data generation to ensure shorter runtimes ( #6948 ) @julian-fong
[ENH] interface to Chronos (zero-shot) foundation model forecaster ( #7001 ) @Z-Fran , @geetu040 , @benHeid , @rigvedmanoj
[ENH] Added get_test_params in ThetaLinesTransformer ( #7199 ) @Anuragwagh
[ENH] cached timesfm instance in case of repeated use ( #7204 ) @fkiraly
[ENH] refactor probabilistic prediction default dispatching logic to mixin class ( #7230 ) @fkiraly
[ENH] TimesFMForecaster dependencies now depends on use_source_package ( #7212 ) @Prtm2110
[ENH] python 3.13 compatibility - SquaringResiduals ( #7244 ) @fkiraly
[ENH] python 3.13 compatibility - craft ( #7251 ) @fkiraly
[ENH] GRU-based time series classifiers ( #6952 ) @fnhirwa
[ENH] HOG1D Transformer add new test parameter set to get_test_params ( #7183 ) @Humorloos
[ENH] Added second parameters example to get_test_params of Filter ( #7178 ) @KarlKolibri
[ENH] temporal radial basis function feature transformer ( #7261 ) @phoeenniixx
[ENH] get_test_params cases in composites conditional on soft dependencies to use _check_estimator_deps ( #7225 ) @shivanshsinghal-22
[DOC] Added reference of load-tecator ( #7171 ) @vedantag17
[DOC] Add example to HOG1DTransformer ( #7180 ) @Humorloos
[DOC] add docstring example for PaddingTransformer ( #7179 ) @jan-mue
[DOC] Added docstring example for HurstExponentTransformer ( #7185 ) @Dehelaan
[DOC] add docstring example to Filter transformer ( #7175 ) @fkiraly
[DOC] clearer install instructions ( #7206 ) @fkiraly
[DOC] Added MOIRAIforecaster to the API docs, create new section on pretrained FM ( #7223 ) @Dehelaan
[DOC] improved docstrings for k-nearest neighbours classifier and regressor ( #7241 ) @fkiraly
[DOC] SlidingWindowSplitter - proper mathematical description ( #7195 ) @fkiraly
[DOC] anomaly and changepoint detection notebook from ODSC 2024 ( #7284 ) @alex-jg3 , @fkiraly
[DOC] Fixes Time Series Segmentation with sktime and ClaSP notebook example ( #7283 ) @RobotPsychologist
[DOC] Fixed minor grammatical errors in README.md ( #7262 ) @Smoothengineer
[DOC] added missing import statements in ReverseAugmenter docstring example ( #7265 ) @sanskarmodi8
[DOC] clarify how fh in forecasting methods is interpreted ( #7227 ) @ericjb
[MNT] Removes coverage upload steps from CI ( #7012 ) @Prtm2110
[MNT] Dependabot: Update skpro requirement from <2.6.0,>=2 to >=2,<2.8.0 ( #7245 ) @dependabot[bot]
[MNT] Dependabot: Update sphinx-gallery requirement from <0.18.0 to <0.19.0 ( #7264 ) @dependabot[bot]
[MNT] Dependabot: Update u8darts requirement from <0.31,>=0.29.0 to >=0.29.0,<0.32 ( #7272 ) @dependabot[bot]
[MNT] Dependabot: Update dask requirement from <2024.8.1 to <2024.10.1 ( #7292 ) @dependabot[bot]
[BUG] Fix ExpandingCutoffSplitter when step_length > 1 ( #7290 ) @ninedigits
[BUG] skip the test load_fpp3 if rdata is not installed ( #7181 ) @jan-mue
[BUG] fixed dask.dataframe import error in dask_to_pd module ( #7260 ) @sanskarmodi8
[BUG] fix ForecastingHorizon.to_absolute with multiple frequencies ( #7172 ) @tianjiqx
[BUG] in ReconcilerForecaster , fix return_totals parameter for all strategies ( #7208 ) @SaiRevanth25
[BUG] Fix DartsLinearRegression failing instead of giving a warning ( #7235 ) @fnhirwa
[BUG] Fix Prophet not tagged as handles-missing-data ( #7267 ) @Garve
[BUG] fix singular matrices in test_reconcilerforecaster_return_totals ( #7293 ) @fkiraly
@alex-jg3 , @Anuragwagh , @benHeid , @Dehelaan , @ericjb , @fkiraly , @fnhirwa , @Garve , @geetu040 , @Humorloos , @jan-mue , @julian-fong , @KarlKolibri , @MarkusSagen , @ninedigits , @phoeenniixx , @Prtm2110 , @RobotPsychologist , @rigvedmanoj , @SaiRevanth25 , @sanskarmodi8 , @shivanshsinghal-22 , @Smoothengineer , @talat-khattatov , @tianjiqx , @vedantag17 , @XinyuWuu , @Z-Fran
new forecaster: interface to Chronos (zero-shot) foundation model forecaster (7001) Z-Fran, geetu040, benHeid, rigvedmanoj
new classifiers: GRU-based time series classifiers (6952) fnhirwa
new transformer: temporal radial basis function features (7261) phoeenniixx
new ExpandingSlidingWindowSplitter, switching from expanding window to sliding window at cutoff point (7193) MarkusSagen
pytorch-forecasting and neuralforecast models now provide probabilistic forecasts for global forecasting (6628, 6666) XinyuWuu
PolynomialTrendForecaster now can make probabilistic forecasts (6424) ericjb
TimesFMForecaster now allows to select the source package, and zero-shot usage has been memory optimized (7204, 7212) Prtm2110, fkiraly
skpro (forecasting soft dependency) bounds have been updated to >=2,<2.8.0
u8darts (forecasting soft dependency) bounds have been updated to >=0.29.0,<0.32.0
dask (data container and parallelization back-end soft dependency) bounds have been updated to <2024.10.1
Performance metrics for probabilistic forecasting, for interval and quantile returns, can now optionally be constructed with an alpha (quantile forecasts) or coverage (interval forecasts) parameter, for example EmpiricalCoverage(coverage=0.7) for "empirical coverage at nominal interval coverage 70 percent".
These are for use with benchmarking or tuning interfaces, where a metric is provided, but predict_interval or predict_quantiles are not explicitly called.
In such a case, the parameter of the metric will be used by the tuner or benchmark utility to tune with the metric at that quantile alpha or interval coverage.
For example, ForecastingGridSearchCV(fcst, cv, scoring=PinballLoss(alpha=[0.3, 0.7])), to tune the forecaster fcst scored by the pinball loss at quantiles 0.3 and 0.7.
[ENH] version tag for objects and estimators (3629) fkiraly
[ENH] add test that html repr of objects does not crash (7151) fkiraly
[ENH] decorators for singleton and multiton oop pattern (7203) fkiraly
[ENH] ExpandingSlidingWindowSplitter, switching from expanding window to sliding window at cutoff point (7193) MarkusSagen
[ENH] added coverage parameter to all metrics of pred_interval type (7278) talat-khattatov
[ENH] refactor datatypes module to scikit-base classes and data records (7161) fkiraly
[ENH] Fix polars PerformanceWarning when object is of type polars.LazyFrame (7221) shivanshsinghal-22
[ENH] Added method predict_interval to PolynomialTrendForecaster (6424) ericjb
[ENH] Quantile Forecast for pytorch-forecasting Models with Global Forecast API (6628) XinyuWuu
[ENH] Global Forecast API for NeuralForecast interface (6666) XinyuWuu
[ENH] testing global forecasters: reduces number of obs in _make_hierarchical for test data generation to ensure shorter runtimes (6948) julian-fong
[ENH] interface to Chronos (zero-shot) foundation model forecaster (7001) Z-Fran, geetu040, benHeid, rigvedmanoj
[ENH] Added get_test_params in ThetaLinesTransformer (7199) Anuragwagh
[ENH] cached timesfm instance in case of repeated use (7204) fkiraly
[ENH] refactor probabilistic prediction default dispatching logic to mixin class (7230) fkiraly
[ENH] TimesFMForecaster dependencies now depends on use_source_package (7212) Prtm2110
[ENH] python 3.13 compatibility - SquaringResiduals (7244) fkiraly
[ENH] python 3.13 compatibility - craft (7251) fkiraly
[ENH] GRU-based time series classifiers (6952) fnhirwa
[ENH] HOG1D Transformer add new test parameter set to get_test_params (7183) Humorloos
[ENH] Added second parameters example to get_test_params of Filter (7178) KarlKolibri
[ENH] temporal radial basis function feature transformer (7261) phoeenniixx
[ENH] get_test_params cases in composites conditional on soft dependencies to use _check_estimator_deps (7225) shivanshsinghal-22
[DOC] Added reference of load-tecator (7171) vedantag17
[DOC] Add example to HOG1DTransformer (7180) Humorloos
[DOC] add docstring example for PaddingTransformer (7179) jan-mue
[DOC] Added docstring example for HurstExponentTransformer (7185) Dehelaan
[DOC] add docstring example to Filter transformer (7175) fkiraly
[DOC] clearer install instructions (7206) fkiraly
[DOC] Added MOIRAIforecaster to the API docs, create new section on pretrained FM (7223) Dehelaan
[DOC] improved docstrings for k-nearest neighbours classifier and regressor (7241) fkiraly
[DOC] SlidingWindowSplitter - proper mathematical description (7195) fkiraly
[DOC] anomaly and changepoint detection notebook from ODSC 2024 (7284) alex-jg3, fkiraly
[DOC] Fixes Time Series Segmentation with sktime and ClaSP notebook example (7283) RobotPsychologist
[DOC] Fixed minor grammatical errors in README.md (7262) Smoothengineer
[DOC] added missing import statements in ReverseAugmenter docstring example (7265) sanskarmodi8
[DOC] clarify how fh in forecasting methods is interpreted (7227) ericjb
[MNT] Removes coverage upload steps from CI (7012) Prtm2110
[MNT] [Dependabot](deps): Update skpro requirement from <2.6.0,>=2 to >=2,<2.8.0 (7245) dependabot[bot]
[MNT] [Dependabot](deps): Update sphinx-gallery requirement from <0.18.0 to <0.19.0 (7264) dependabot[bot]
[MNT] [Dependabot](deps): Update u8darts requirement from <0.31,>=0.29.0 to >=0.29.0,<0.32 (7272) dependabot[bot]
[MNT] [Dependabot](deps): Update dask requirement from <2024.8.1 to <2024.10.1 (7292) dependabot[bot]
[BUG] Fix ExpandingCutoffSplitter when step_length > 1 (7290) ninedigits
[BUG] skip the test load_fpp3 if rdata is not installed (7181) jan-mue
[BUG] fixed dask.dataframe import error in dask_to_pd module (7260) sanskarmodi8
[BUG] fix ForecastingHorizon.to_absolute with multiple frequencies (7172) tianjiqx
[BUG] in ReconcilerForecaster, fix return_totals parameter for all strategies (7208) SaiRevanth25
[BUG] Fix DartsLinearRegression failing instead of giving a warning (7235) fnhirwa
[BUG] Fix Prophet not tagged as handles-missing-data (7267) Garve
[BUG] fix singular matrices in test_reconcilerforecaster_return_totals (7293) fkiraly
alex-jg3, Anuragwagh, benHeid, Dehelaan, ericjb, fkiraly, fnhirwa, Garve, geetu040, Humorloos, jan-mue, julian-fong, KarlKolibri, MarkusSagen, ninedigits, phoeenniixx, Prtm2110, RobotPsychologist, rigvedmanoj, SaiRevanth25, sanskarmodi8, shivanshsinghal-22, Smoothengineer, talat-khattatov, tianjiqx, vedantag17, XinyuWuu, Z-Fran
Please see our changelog for a description of all changes.
Feature release.
Please see our changelog for a description of all changes.
@ankit-1204, @benHeid, @ericjb, @fkiraly, @pranavvp16, @SaiRevanth25, @Saptarshi-Bandopadhyay, @XinyuWuu, @yarnabrina
Full Changelog: https://github.com/sktime/sktime/compare/v0.33.0...v0.33.1
Small feature release for showcase at pydata Paris.
interface for MOIRAI foundation model ( #6746 ) @pranavvp16 , @benHeid
GroupbyCategoryForecaster for applying panel forecasting by category or segment ( #7066 ) @felipeangelimvieira
In ReconcilerForecaster , users can now choose to return a dataframe without __total ( #7127 ) @SaiRevanth25
time series segmentation via clustering ( #6782 ) @ankit-1204
logger transformer for logging pipeline inputs and outputs ( #7074 ) @fkiraly
optuna (hyperparameter optimization soft dependency) bounds have been updated to <4.1
mne (transformations soft dependency) bounds have been updated to >=1.5,<1.9
[ENH] is_scitype utility for scitype checking, improve support of estimators with multiple object types ( #7143 ) @benHeid
[ENH] Hierarchical scitype support for polars ( #6697 ) @pranavvp16
[ENH] refactor datatypes example fixtures to BaseObject classes ( #7133 ) @fkiraly
[ENH] Add interface for MOIRAI foundation model ( #6746 ) @pranavvp16 , @benHeid
[ENH] ThetaForecaster - add one more parameter set ( #7055 ) @fkiraly
[ENH] GroupbyCategoryForecaster for applying panel forecasting by category or segment ( #7066 ) @felipeangelimvieira
[ENH] Adding a parameter to ReconcilerForecaster to return a dataframe without the dunder levels ( #7127 ) @SaiRevanth25
[ENH] add _predict_var in test_pred_int_tag ( #7154 ) @fkiraly
[ENH] retrieval utilities for all functions or classes in a module ( #7089 ) @fkiraly
[ENH] time series segmentation via clustering ( #6782 ) @ankit-1204
[ENH] logger transformer for logging pipeline inputs and outputs ( #7074 ) @fkiraly
[DOC] correct incorrect bounds mentioned in 0.33.0 changelog ( #7114 ) @fkiraly
[DOC] Added a docstring example to TimeSeriesKernelKMeans ( #7124 ) @Saptarshi-Bandopadhyay
[DOC] update core developers on team page, formatting ( #7140 ) @fkiraly
[DOC] minor documentation fixes ( #7141 ) @fkiraly
[DOC] add quicklinks at top of README box ( #7159 ) @fkiraly
[DOC] remove 2024 elections registration news item ( #7158 ) @fkiraly
[DOC] remove references to nested_univ from extension templates ( #7058 ) @fkiraly
[MNT] Updating pre-commit hooks and corresponding changes ( #7109 ) @yarnabrina
[MNT] Dependabot: Update optuna requirement from <3.7 to <4.1 ( #7104 ) @dependabot[bot]
[MNT] Dependabot: Update mne requirement from <1.8,>=1.5 to >=1.5,<1.9 ( #7129 ) @dependabot[bot]
[MNT] temporarily skip temporian related failure of test_complex_function until #7040 is resolved ( #7147 ) @fkiraly
[MNT] fix FPP3 download link ( #7164 ) @fkiraly , @ericjb
[MNT] bound SignatureTransformer to numpy<2 ( #7163 ) @fkiraly
[MNT] remove ptf install from example notebooks ( #7165 ) @XinyuWuu
[MNT] remove python version bound from pytorch-forecasting based estimators ( #7102 ) @fkiraly
[BUG] Fix pykan forecaster ( #7150 ) @benHeid
[BUG] fix probabilistic forecasts if only _predict_var is implemented ( #7153 ) @fkiraly
[BUG] fix drop_na and update mode of Differencer transformation ( #7115 ) @fkiraly
Test framework
@ankit-1204 , @benHeid , @ericjb , @fkiraly , @pranavvp16 , @SaiRevanth25 , @Saptarshi-Bandopadhyay , @XinyuWuu , @yarnabrina
Small feature release for showcase at pydata Paris.
interface for MOIRAI foundation model (6746) pranavvp16, benHeid
GroupbyCategoryForecaster for applying panel forecasting by category or segment (7066) felipeangelimvieira
In ReconcilerForecaster, users can now choose to return a dataframe without __total (7127) SaiRevanth25
time series segmentation via clustering (6782) ankit-1204
logger transformer for logging pipeline inputs and outputs (7074) fkiraly
optuna (hyperparameter optimization soft dependency) bounds have been updated to <4.1
mne (transformations soft dependency) bounds have been updated to >=1.5,<1.9
[ENH] is_scitype utility for scitype checking, improve support of estimators with multiple object types (7143) benHeid
[ENH] Hierarchical scitype support for polars (6697) pranavvp16
[ENH] refactor datatypes example fixtures to BaseObject classes (7133) fkiraly
[ENH] Add interface for MOIRAI foundation model (6746) pranavvp16, benHeid
[ENH] ThetaForecaster - add one more parameter set (7055) fkiraly
[ENH] GroupbyCategoryForecaster for applying panel forecasting by category or segment (7066) felipeangelimvieira
[ENH] Adding a parameter to ReconcilerForecaster to return a dataframe without the dunder levels (7127) SaiRevanth25
[ENH] add _predict_var in test_pred_int_tag (7154) fkiraly
[ENH] retrieval utilities for all functions or classes in a module (7089) fkiraly
[ENH] time series segmentation via clustering (6782) ankit-1204
[ENH] logger transformer for logging pipeline inputs and outputs (7074) fkiraly
[DOC] correct incorrect bounds mentioned in 0.33.0 changelog (7114) fkiraly
[DOC] Added a docstring example to TimeSeriesKernelKMeans (7124) Saptarshi-Bandopadhyay
[DOC] update core developers on team page, formatting (7140) fkiraly
[DOC] minor documentation fixes (7141) fkiraly
[DOC] add quicklinks at top of README box (7159) fkiraly
[DOC] remove 2024 elections registration news item (7158) fkiraly
[DOC] remove references to nested_univ from extension templates (7058) fkiraly
[MNT] Updating pre-commit hooks and corresponding changes (7109) yarnabrina
[MNT] [Dependabot](deps): Update optuna requirement from <3.7 to <4.1 (7104) dependabot[bot]
[MNT] [Dependabot](deps): Update mne requirement from <1.8,>=1.5 to >=1.5,<1.9 (7129) dependabot[bot]
[MNT] temporarily skip temporian related failure of test_complex_function until #7040 is resolved (7147) fkiraly
[MNT] fix FPP3 download link (7164) fkiraly, ericjb
[MNT] bound SignatureTransformer to numpy<2 (7163) fkiraly
[MNT] remove ptf install from example notebooks (7165) XinyuWuu
[MNT] remove python version bound from pytorch-forecasting based estimators (7102) fkiraly
[BUG] Fix pykan forecaster (7150) benHeid
[BUG] fix probabilistic forecasts if only _predict_var is implemented (7153) fkiraly
[BUG] fix drop_na and update mode of Differencer transformation (7115) fkiraly
Test framework
ankit-1204, benHeid, ericjb, fkiraly, pranavvp16, SaiRevanth25, Saptarshi-Bandopadhyay, XinyuWuu, yarnabrina
Maintenance release with scheduled deprecations and change actions. For last larger feature updates, see 0.32.4 and 0.32.2
Maintenance release with scheduled deprecations and change actions. For last larger feature updates, see 0.32.4 and 0.32.2
Please see our changelog for a description of all changes.
@benHeid, @ericjb, @fkiraly, @SaiRevanth25, @Saptarshi-Bandopadhyay
Full Changelog: https://github.com/sktime/sktime/compare/v0.32.4...v0.33.0
Maintenance release, with scheduled deprecations and change actions.
For last non-maintenance content updates, see 0.32.4 and 0.32.2.
scikit-base (core dependency) bounds have been updated to >=0.6.1,<0.10.0
holidays (transformations soft dependency) bounds have been updated to >=0.29,<0.57
pykan (deep learning soft dependency) bounds have been updated to >=0.2,<0.2.7
mne (transformations soft dependency) bounds have been updated to >=1.5,<1.9
dask (data container and parallelization back-end soft dependency) bounds have been updated to <2024.8.3
pytorch-forecasting (forecasting soft dependency) bounds have been updated to >=1.0.0,<1.2.0
in DirectReductionForecaster the default for windows_identical has changed to False .
[MNT] Try to reduce load for the runners ( #7061 ) @benHeid
[MNT] 0.33.0 deprecations and change actions ( #7091 ) @fkiraly
[MNT] ffp3 datasets URLs changed on CRAN; updated _fpp3_loaders.py accordingly ( #7084 ) @ericjb
[MNT] remove <3.11 restriction for pytorch-forecasting , add upper bound ( #7092 ) @fkiraly
[MNT] Dependabot: Update dask requirement from <2024.8.2 to <2024.8.3 ( #7062 ) @dependabot[bot]
[MNT] Dependabot: Update numpy requirement from <2.1,>=1.21 to >=1.21,<2.2 ( #7007 ) @dependabot[bot]
[MNT] Dependabot: Update scikit-base requirement from <0.9.0,>=0.6.1 to >=0.6.1,<0.10.0 ( #7035 ) @dependabot[bot]
[MNT] Dependabot: Update holidays requirement from <0.56,>=0.29 to >=0.29,<0.57 ( #7096 ) @dependabot[bot]
[MNT] Dependabot: Update pykan requirement from <0.2.2,>=0.2 to >=0.2,<0.2.7 ( #7010 ) @dependabot[bot]
[MNT] Dependabot: Update mne requirement from <1.8,>=1.5 to >=1.5,<1.9 ( #7004 ) @dependabot[bot]
[DOC] Adds @SaiRevanth25` contributions to all-contributors file ( #7085 ) @SaiRevanth25
[DOC] fix typo and formatting in installation docs ( #7060 ) @Saptarshi-Bandopadhyay
[ENH] change test_inheritance to be more lenient to framework level extensions ( #7067 ) @fkiraly
@benHeid , @ericjb , @fkiraly , @SaiRevanth25 , @Saptarshi-Bandopadhyay
Maintenance release, with scheduled deprecations and change actions.
For last non-maintenance content updates, see 0.32.4 and 0.32.2.
scikit-base (core dependency) bounds have been updated to >=0.6.1,<0.10.0
holidays (transformations soft dependency) bounds have been updated to >=0.29,<0.57
pykan (deep learning soft dependency) bounds have been updated to >=0.2,<0.2.7
mne (transformations soft dependency) bounds have been updated to >=1.5,<1.9
dask (data container and parallelization back-end soft dependency) bounds have been updated to <2024.8.3
pytorch-forecasting (forecasting soft dependency) bounds have been updated to >=1.0.0,<1.2.0
in DirectReductionForecaster the default for windows_identical has changed to False.
[MNT] Try to reduce load for the runners (7061) benHeid
[MNT] 0.33.0 deprecations and change actions (7091) fkiraly
[MNT] ffp3 datasets URLs changed on CRAN; updated _fpp3_loaders.py accordingly (7084) ericjb
[MNT] remove <3.11 restriction for pytorch-forecasting, add upper bound (7092) fkiraly
[MNT] [Dependabot](deps): Update dask requirement from <2024.8.2 to <2024.8.3 (7062) dependabot[bot]
[MNT] [Dependabot](deps): Update numpy requirement from <2.1,>=1.21 to >=1.21,<2.2 (7007) dependabot[bot]
[MNT] [Dependabot](deps): Update scikit-base requirement from <0.9.0,>=0.6.1 to >=0.6.1,<0.10.0 (7035) dependabot[bot]
[MNT] [Dependabot](deps): Update holidays requirement from <0.56,>=0.29 to >=0.29,<0.57 (7096) dependabot[bot]
[MNT] [Dependabot](deps): Update pykan requirement from <0.2.2,>=0.2 to >=0.2,<0.2.7 (7010) dependabot[bot]
[MNT] [Dependabot](deps): Update mne requirement from <1.8,>=1.5 to >=1.5,<1.9 (7004) dependabot[bot]
[DOC] Adds SaiRevanth25` contributions to all-contributors file (7085) SaiRevanth25
[DOC] fix typo and formatting in installation docs (7060) Saptarshi-Bandopadhyay
[ENH] change test_inheritance to be more lenient to framework level extensions (7067) fkiraly
benHeid, ericjb, fkiraly, SaiRevanth25, Saptarshi-Bandopadhyay
Hotfix release for colab compatibility, and minor feature release.
Hotfix release for colab compatibility, and minor feature release.
Please see our changelog for a description of all changes.
@Alex-JG3, @fkiraly, @fnhirwa, @geetu040, @ninedigits, @phoeenniixx, @Saptarshi-Bandopadhyay, @wirrywoo
Full Changelog: https://github.com/sktime/sktime/compare/v0.32.3...v0.32.4
Small feature and fix release for:
colab compatibility, hotfix for polars dependency incompatibility
compatibility with skchange 2nd party integration
improvements to the anomalies, changepoints, segmentation framework
documentation update related to upcoming sktime elections
A scitype adaptation framework is introduced, allowing use of an estimator of one type in a slot for another, via type coercion. The coercion framework covers sklearn , skpro , and sktime estimators currently, and is extensible. A user and developer shorthand for such coercion is provided as the registry.coerce_scitype function, which dispatches to individual coercion mechanisms.
Time series clusterers that produce a cluster assignment can be coerced to transformations, enabling their use in any pipeline slot for transformations.
Anomaly and changepoint detectors can now be pipelined with transformations, resulting in a detector. Dunder concatenation transformer * detector will default to this.
Anomaly and changepoint detectors can be coerced to transformations, enabling their use in any pipeline slot for transformations.
[ENH] decouple registry from base modules, scitype specific data records for documentation of estimator types ( #6998 ) @fkiraly
[ENH] scitype coercion and checking utility ( #6969 ) @fkiraly
[ENH] Add padded f1 score for evaluating change point detection algorithms ( #7034 ) @Alex-JG3
[ENH] improvements to BaseSeriesAnnotator base class for anomaly, changepoint, segments ( #7073 ) @fkiraly
[ENH] pipeline for anomaly, changepoint detectors and segmenters ( #7071 ) @fkiraly
[ENH] coercion to use time series anomaly, changepoint detectors as transformers ( #7072 ) @fkiraly
[ENH] enable use of clusterers as transformations, enable TransformSelectForecaster use of clusterers for group selection ( #7068 ) @fkiraly
[ENH] Hurst exponent feature extraction transformer ( #7065 ) @phoeenniixx
[DOC] Add SVG version of the sktime logo with no text ( #7024 ) @wirrywoo
[DOC] Improve documentation for TinyTimeMixer ( #7009 ) @geetu040
[DOC] fix broken links in continuous integration docs ( #7059 ) @Saptarshi-Bandopadhyay
[DOC] split list of transformation pipeline components into subcategories ( #7075 ) @fkiraly
[BUG] fix sktime crash with older polars versions ( #7057 ) @fkiraly
[BUG] Fix ForecastX.update when the forecaster_X_exogeneous is set to "complement" ( #7041 ) @fnhirwa
[BUG] fix ExpandingCutoffSplitter for case where fh is not continuous ( #7053 ) @ninedigits
@Alex-JG3 , @fkiraly , @fnhirwa , @geetu040 , @ninedigits , @phoeenniixx , @Saptarshi-Bandopadhyay , @wirrywoo
Small feature and fix release for:
colab compatibility, hotfix for polars dependency incompatibility
compatibility with skchange 2nd party integration
improvements to the anomalies, changepoints, segmentation framework
documentation update related to upcoming sktime elections
A scitype adaptation framework is introduced, allowing use of an estimator of one type in a slot for another, via type coercion. The coercion framework covers sklearn, skpro, and sktime estimators currently, and is extensible. A user and developer shorthand for such coercion is provided as the registry.coerce_scitype function, which dispatches to individual coercion mechanisms.
Time series clusterers that produce a cluster assignment can be coerced to transformations, enabling their use in any pipeline slot for transformations.
Anomaly and changepoint detectors can now be pipelined with transformations, resulting in a detector. Dunder concatenation transformer * detector will default to this.
Anomaly and changepoint detectors can be coerced to transformations, enabling their use in any pipeline slot for transformations.
[ENH] decouple registry from base modules, scitype specific data records for documentation of estimator types (6998) fkiraly
[ENH] scitype coercion and checking utility (6969) fkiraly
[ENH] Add padded f1 score for evaluating change point detection algorithms (7034) Alex-JG3
[ENH] improvements to BaseSeriesAnnotator base class for anomaly, changepoint, segments (7073) fkiraly
[ENH] pipeline for anomaly, changepoint detectors and segmenters (7071) fkiraly
[ENH] coercion to use time series anomaly, changepoint detectors as transformers (7072) fkiraly
[ENH] enable use of clusterers as transformations, enable TransformSelectForecaster use of clusterers for group selection (7068) fkiraly
[ENH] Hurst exponent feature extraction transformer (7065) phoeenniixx
[DOC] Add SVG version of the sktime logo with no text (7024) wirrywoo
[DOC] Improve documentation for TinyTimeMixer (7009) geetu040
[DOC] fix broken links in continuous integration docs (7059) Saptarshi-Bandopadhyay
[DOC] split list of transformation pipeline components into subcategories (7075) fkiraly
[BUG] fix sktime crash with older polars versions (7057) fkiraly
[BUG] Fix ForecastX.update when the forecaster_X_exogeneous is set to "complement" (7041) fnhirwa
[BUG] fix ExpandingCutoffSplitter for case where fh is not continuous (7053) ninedigits
Alex-JG3, fkiraly, fnhirwa, geetu040, ninedigits, phoeenniixx, Saptarshi-Bandopadhyay, wirrywoo
Hotfix: bugfix for html representation of forecasting pipelines.
Hotfix: bugfix for html representation of forecasting pipelines.
Full Changelog: https://github.com/sktime/sktime/compare/v0.32.2...v0.32.3
Hotfix release with bugfix for html representation of forecasting pipelines.
For last non-maintenance content updates, see 0.32.2.
[BUG] fix html display for TransformedTargetForecaster and ForecastingPipeline
Please see our changelog for a description of all changes.
Feature release.
Please see our changelog for a description of all changes.
@Abhay-Lejith, @felipeangelimvieira, @fkiraly, @geetu040, @marrov, @meraldoantonio, @ninedigits, @pranavvp16, @SaiRevanth25, @shlok191, @toandaominh1997, @wirrywoo, @wpdonders, @yarnabrina
Full Changelog: https://github.com/sktime/sktime/compare/v0.32.1...v0.32.2
HierarchicalProphet forecaster from prophetverse ( #7028 ) @felipeangelimvieira
Regularized VAR reduction forecaster, VARReduce ( #6725 ) @meraldoantonio
Interface to TimesFM Forecaster ( #6571 ) @geetu040
Subsequence Extraction Transformer ( #6967 ) @wirrywoo
Framework support for categorical data has been extended to transformers and pipelines ( #6924 ) @Abhay-Lejith
Clusterer tags for capability to assign cluster centers ( #7018 ) @fkiraly
holiday (transformations soft dependency) bounds have been updated to >=0.29,<0.56
dask (data container and parallelization back-end) bounds have been updated to <2024.8.2
New tags for clusterers have been added to characterize capabilities to assign cluster centers. The following boolean tags have been added:
capability:predict , whether the clusterer can assign cluster labels via predict
capability:predict_proba , for probabilistic cluster assignment
capability:out_of_sample , for out-of-sample cluster assignment. If False, the clusterer can only assign clusters to data points seen during fitting.
[ENH] placeholder record decorator ( #7029 ) @fkiraly
[ENH] Hierarchical sales toydata generator from workshops ( #6953 ) @marrov
[ENH] Convert the date column to a period with daily frequency in load_m5 ( #6990 ) @SaiRevanth25
[ENH] Polars Series scitype supports ( #6485 ) @pranavvp16
[ENH] Polars Panel scitype support ( #6552 ) @pranavvp16
[ENH] Addition of feature_kind metadata attribute to gluonts datatypes ( #6871 ) @shlok191
[ENH] interface to TimesFM Forecaster ( #6571 ) @geetu040
[ENH] New regularized VAR reduction forecaster, VARReduce ( #6725 ) @meraldoantonio
[ENH] Add HierarchicalProphet estimator to prophetverse module ( #7028 ) @felipeangelimvieira
[ENH] clusterer tags for capability to assign cluster centers ( #7018 ) @fkiraly
[ENH] Extending categorical support in X to transformers and pipelines ( #6924 ) @Abhay-Lejith
[ENH] Subsequence Extraction Transformer ( #6967 ) @wirrywoo
[DOC] minor improvements to docstring of Bollinger (bands) ( #6978 ) @fkiraly
[DOC] Update .all-contributorsrc with council roles ( #6962 ) @fkiraly
[DOC] update soft dependency handling guide for estimators ( #7000 ) @fkiraly
[DOC] improvements to docstrings for panel tasks - time series classification, regression, clustering ( #6991 ) @fkiraly
[DOC] update XinyuWuu’s user name ( #7030 ) @fkiraly
[DOC] fixes to TransformedTargetForecaster docstring ( #7002 ) @fkiraly
[DOC] update intro notebook with material from ISF and EuroSciPy 2024 ( #7013 ) @fkiraly
[DOC] Fix docstring for ExpandingCutoffSplitter ( #7033 ) @ninedigits
[DOC] fix incorrect import in EnbPIForecaster docstring ( #7015 ) @fkiraly
[MNT] Refactor show_versions to use dependencies module ( #6883 ) @fkiraly
[MNT] sync changelog with hotfix branch anirban-sktime-0.31.2 ( #6963 ) @yarnabrina
[MNT] add numpy 2 incompatibility flag to pmdarima dependency ( #6974 ) @fkiraly
[MNT] decorate test_auto_arima with numpy 2 skip until final fix/diagnosis ( #6973 ) @fkiraly
[MNT] remove tsbootstrap dependency from public dependency sets ( #6966 ) @fkiraly
[MNT] rename base class TimeSeriesLloyds to BaseTimeSeriesLloyds ( #6992 ) @fkiraly
[MNT] remove module level numba import warnings ( #6999 ) @fkiraly
[MNT] esig based estimators: add numpy<2 bound ( #7036 ) @fkiraly
[MNT] Dependabot: Bump tj-actions/changed-files from 44 to 45 ( #7019 ) @dependabot[bot]
[MNT] Dependabot: Update holidays requirement from <0.55,>=0.29 to >=0.29,<0.56 ( #7006 ) @dependabot[bot]
[MNT] Dependabot: Update dask requirement from <2024.8.1 to <2024.8.2 ( #7005 ) @dependabot[bot]
[BUG] fix test_softdep_error dependency handling check if environment marker tag is not satisfied ( #6961 ) @fkiraly
[BUG] fix dependency checkers in case of multiple distributions available in environment, e.g., on databricks ( #6986 ) @fkiraly , @toandaominh1997
[BUG] Fix ForecastingBenchmark giving an error when the dataloader returns the tuple (y, X) ( #6971 ) @SaiRevanth25
[BUG] Fix nested_univ converter inconsistent handling of index level names ( #7026 ) @pranavvp16
[BUG] TinyTimeMixerForecaster : fix truncating index and update test_params ( #6965 ) @geetu040
[BUG] Do not add season condition names as extra regressors in Prophet ( #6988 ) @wpdonders
[BUG] Fix Prophet _get_fitted_params error when the timeseries is constant ( #7011 ) @felipeangelimvieira
@Abhay-Lejith , @felipeangelimvieira , @fkiraly , @geetu040 , @marrov , @meraldoantonio , @ninedigits , @pranavvp16 , @SaiRevanth25 , @shlok191 , @toandaominh1997 , @wirrywoo , @wpdonders , @yarnabrina
HierarchicalProphet forecaster from prophetverse (7028) felipeangelimvieira
Regularized VAR reduction forecaster, VARReduce (6725) meraldoantonio
Interface to TimesFM Forecaster (6571) geetu040
Subsequence Extraction Transformer (6967) wirrywoo
Framework support for categorical data has been extended to transformers and pipelines (6924) Abhay-Lejith
Clusterer tags for capability to assign cluster centers (7018) fkiraly
holiday (transformations soft dependency) bounds have been updated to >=0.29,<0.56
dask (data container and parallelization back-end) bounds have been updated to <2024.8.2
New tags for clusterers have been added to characterize capabilities to assign cluster centers. The following boolean tags have been added:
capability:predict, whether the clusterer can assign cluster labels via predict
capability:predict_proba, for probabilistic cluster assignment
capability:out_of_sample, for out-of-sample cluster assignment. If False, the clusterer can only assign clusters to data points seen during fitting.
[ENH] placeholder record decorator (7029) fkiraly
[ENH] Hierarchical sales toydata generator from workshops (6953) marrov
[ENH] Convert the date column to a period with daily frequency in load_m5 (6990) SaiRevanth25
[ENH] Polars Series scitype supports (6485) pranavvp16
[ENH] Polars Panel scitype support (6552) pranavvp16
[ENH] Addition of feature_kind metadata attribute to gluonts datatypes (6871) shlok191
[ENH] interface to TimesFM Forecaster (6571) geetu040
[ENH] New regularized VAR reduction forecaster, VARReduce (6725) meraldoantonio
[ENH] Add HierarchicalProphet estimator to prophetverse module (7028) felipeangelimvieira
[ENH] clusterer tags for capability to assign cluster centers (7018) fkiraly
[ENH] Extending categorical support in X to transformers and pipelines (6924) Abhay-Lejith
[ENH] Subsequence Extraction Transformer (6967) wirrywoo
[DOC] minor improvements to docstring of Bollinger (bands) (6978) fkiraly
[DOC] Update .all-contributorsrc with council roles (6962) fkiraly
[DOC] update soft dependency handling guide for estimators (7000) fkiraly
[DOC] improvements to docstrings for panel tasks - time series classification, regression, clustering (6991) fkiraly
[DOC] update XinyuWuu's user name (7030) fkiraly
[DOC] fixes to TransformedTargetForecaster docstring (7002) fkiraly
[DOC] update intro notebook with material from ISF and EuroSciPy 2024 (7013) fkiraly
[DOC] Fix docstring for ExpandingCutoffSplitter (7033) ninedigits
[DOC] fix incorrect import in EnbPIForecaster docstring (7015) fkiraly
[MNT] Refactor show_versions to use dependencies module (6883) fkiraly
[MNT] sync changelog with hotfix branch anirban-sktime-0.31.2 (6963) yarnabrina
[MNT] add numpy 2 incompatibility flag to pmdarima dependency (6974) fkiraly
[MNT] decorate test_auto_arima with numpy 2 skip until final fix/diagnosis (6973) fkiraly
[MNT] remove tsbootstrap dependency from public dependency sets (6966) fkiraly
[MNT] rename base class TimeSeriesLloyds to BaseTimeSeriesLloyds (6992) fkiraly
[MNT] remove module level numba import warnings (6999) fkiraly
[MNT] esig based estimators: add numpy<2 bound (7036) fkiraly
[MNT] [Dependabot](deps): Bump tj-actions/changed-files from 44 to 45 (7019) dependabot[bot]
[MNT] [Dependabot](deps): Update holidays requirement from <0.55,>=0.29 to >=0.29,<0.56 (7006) dependabot[bot]
[MNT] [Dependabot](deps): Update dask requirement from <2024.8.1 to <2024.8.2 (7005) dependabot[bot]
[BUG] fix test_softdep_error dependency handling check if environment marker tag is not satisfied (6961) fkiraly
[BUG] fix dependency checkers in case of multiple distributions available in environment, e.g., on databricks (6986) fkiraly, toandaominh1997
[BUG] Fix ForecastingBenchmark giving an error when the dataloader returns the tuple (y, X) (6971) SaiRevanth25
[BUG] Fix nested_univ converter inconsistent handling of index level names (7026) pranavvp16
[BUG] TinyTimeMixerForecaster: fix truncating index and update test_params (6965) geetu040
[BUG] Do not add season condition names as extra regressors in Prophet (6988) wpdonders
[BUG] Fix Prophet ``_get_fitted_params ``error when the timeseries is constant (7011) felipeangelimvieira
Abhay-Lejith, felipeangelimvieira, fkiraly, geetu040, marrov, meraldoantonio, ninedigits, pranavvp16, SaiRevanth25, shlok191, toandaominh1997, wirrywoo, wpdonders, yarnabrina
Hotfix: fix make_reduction type inference fallback for not fully sklearn compliant regression estimators, e.g., catboost.
Hotfix: fix make_reduction type inference fallback for not fully sklearn compliant regression estimators, e.g., catboost.
Full Changelog: https://github.com/sktime/sktime/compare/v0.32.0...v0.32.1
Hotfix release for using make_reduction with not fully sklearn compliant tabular regressors such as from catboost .
For last non-maintenance content updates, see 0.31.1.
[BUG] fix make_reduction type inference for non-sklearn estimators
Maintenance release with scheduled deprecations and change actions. For last larger feature update, see 0.31.1.
Maintenance release with scheduled deprecations and change actions. For last larger feature update, see 0.31.1.
Please see our changelog for a description of all changes.
@fkiraly, @hliebert, @yarnabrina
Full Changelog: https://github.com/sktime/sktime/compare/v0.31.1...v0.32.0
Maintenance release, with scheduled deprecations and change actions.
For last non-maintenance content updates, see 0.31.1.
skpro (soft dependency) bounds have been updated to >=2,<2.6.0
skforecast (forecasting soft dependency) bounds have been updated to <0.14.0 .
all sktime estimators and objects are now required to have at least two test parameter sets in get_test_params to be compliant with check_estimator contract tests. This requirement was previously stated in the extension template but not enforced. It is now also included in the automated tests via check_estimator . Estimators without (unreserved) parameters, i.e., where two distinct parameter sets are not possible, are excepted from this.
From sktime 0.38.0 , forecasters’ predict_proba will require skpro to be present in the python environment, for distribution objects to represent distributional forecasts. Until sktime 0.35.0 , predict_proba will continue working without skpro , defaulting to return objects in sktime.proba if skpro is not present. From sktime 0.35.0 , an error will be raised upon call of forecaster predict_proba if skpro is not present in the environment. Users of forecasters’ predict_proba should ensure that skpro is installed in the environment.
The probability distributions module sktime.proba deprecated and will be fully replaced by skpro in sktime 0.38.0 . Until sktime 0.38.0 , imports from sktime.proba will continue working, defaulting to sktime.proba if skpro is not present, otherwise redirecting imports to skpro objects. From sktime 0.35.0 , an error will be raised if skpro is not present in the environment, otherwise imports are redirected to skpro . Direct or indirect users of sktime.proba should ensure skpro is installed in the environment. Direct users of the sktime.proba module should, in addition, replace any imports from sktime.proba with imports from skpro.distributions .
[MNT] 0.32.0 deprecations and change actions ( #6916 ) @fkiraly
[MNT] Dependabot: Update skpro requirement from <2.5.0,>=2 to >=2,<2.6.0 ( #6897 ) @dependabot[bot]
[MNT] remove numpy 2 incompatibility flag from numba based estimators ( #6915 ) @fkiraly
[MNT] isolate joblib ( #6385 ) @fkiraly
[MNT] handle more pandas deprecations ( #6941 ) @fkiraly
[MNT] deprecation of proba module in favour of skpro soft dependency ( #6940 ) @fkiraly
[MNT] update versions of pre-commit hooks ( #6947 ) @yarnabrina
[MNT] 0.32.0 release action - revert temporary skip get_test_params number check for 0.21.1 and 0.22.0 release ( #5114 ) @fkiraly
[MNT] Bump skforecast to 0.13 version allowing support for python 3.12 ( #6946 ) @yarnabrina
[BUG] Fix Xt_msg type in transformations.base ( #6944 ) @hliebert
@fkiraly , @hliebert , @yarnabrina
Maintenance release, with scheduled deprecations and change actions.
For last non-maintenance content updates, see 0.31.1.
skpro (soft dependency) bounds have been updated to >=2,<2.6.0
skforecast (forecasting soft dependency) bounds have been updated to <0.14.0.
all sktime estimators and objects are now required to have at least two test parameter sets in get_test_params to be compliant with check_estimator contract tests. This requirement was previously stated in the extension template but not enforced. It is now also included in the automated tests via check_estimator. Estimators without (unreserved) parameters, i.e., where two distinct parameter sets are not possible, are excepted from this.
From sktime 0.38.0, forecasters' predict_proba will require skpro to be present in the python environment, for distribution objects to represent distributional forecasts. Until sktime 0.35.0, predict_proba will continue working without skpro, defaulting to return objects in sktime.proba if skpro is not present. From sktime 0.35.0, an error will be raised upon call of forecaster predict_proba if skpro is not present in the environment. Users of forecasters' predict_proba should ensure that skpro is installed in the environment.
The probability distributions module sktime.proba deprecated and will be fully replaced by skpro in sktime 0.38.0. Until sktime 0.38.0, imports from sktime.proba will continue working, defaulting to sktime.proba if skpro is not present, otherwise redirecting imports to skpro objects. From sktime 0.35.0, an error will be raised if skpro is not present in the environment, otherwise imports are redirected to skpro. Direct or indirect users of sktime.proba should ensure skpro is installed in the environment. Direct users of the sktime.proba module should, in addition, replace any imports from sktime.proba with imports from skpro.distributions.
[MNT] 0.32.0 deprecations and change actions (6916) fkiraly
[MNT] [Dependabot](deps): Update skpro requirement from <2.5.0,>=2 to >=2,<2.6.0 (6897) dependabot[bot]
[MNT] remove numpy 2 incompatibility flag from numba based estimators (6915) fkiraly
[MNT] isolate joblib (6385) fkiraly
[MNT] handle more pandas deprecations (6941) fkiraly
[MNT] deprecation of proba module in favour of skpro soft dependency (6940) fkiraly
[MNT] update versions of pre-commit hooks (6947) yarnabrina
[MNT] 0.32.0 release action - revert temporary skip get_test_params number check for 0.21.1 and 0.22.0 release (5114) fkiraly
[MNT] Bump skforecast to 0.13 version allowing support for python 3.12 (6946) yarnabrina
[BUG] Fix Xt_msg type in transformations.base (6944) hliebert
fkiraly, hliebert, yarnabrina
Hotfix: fix make_reduction type inference fallback for not fully sklearn compliant regression estimators, e.g., catboost.
Hotfix: fix make_reduction type inference fallback for not fully sklearn compliant regression estimators, e.g., catboost.
Full Changelog: https://github.com/sktime/sktime/compare/v0.31.1...v0.31.2
Hotfix release, released after hotfix release 0.32.1, to apply the same hotfix to 0.31.X versions as well.
Hotfix for using make_reduction with not fully sklearn compliant tabular regressors such as from catboost .
For last non-maintenance content updates, see 0.31.1.
[BUG] fix make_reduction type inference for non-sklearn estimators
This is a hotfix for 0.31.1 release, fixing a regression. This release is not contained in the 0.32.0 or 0.32.1 releases.
Please see our changelog for a description of all changes.
Feature release.
Please see our changelog for a description of all changes.
@Abhay-Lejith, @Alex-JG3, @alexander-lakocy, @AlexeyOm, @bastisar, @benHeid, @ceroper, @DinoBektesevic, @fkiraly, @fnhirwa, @fr1ll, @gareth-brown-86, @geetu040, @hliebert, @julian-fong, @mateuszkasprowicz, @MBristle, @melinny, @Mitchjkjkjk, @mk406, @SaiRevanth25, @sbhobbes, @shlok191, @SultanOrazbayev, @szepeviktor, @Xinyu-Wu-0000, @yarnabrina, @ZhipengXue97
Full Changelog: https://github.com/sktime/sktime/compare/v0.31.0...v0.31.1
html representation of objects now has a button linking to documentation page ( #6876 ) @mateuszkasprowicz
interface to TinyTimeMixer foundation model ( #6712 ) @geetu040
interface to autots ensemble ( #5948 ) @MBristle
interface to darts reduction models ( #6712 ) @fnhirwa , @yarnabrina
LTSFTransformer based on cure-lab research code base ( #6202 ) @geetu040
MVTS transformer classifier ( #6791 ) @geetu040
forecasters can now support categorical X , as per tag ( #6704 , #6732 ) @Abhay-Lejith
DirectReductionForecaster now has a windows_identical option ( #6650 ) @hliebert
ForecastingOptunaSearchCV can now be passed custom samplers and “higher is better” scores ( #6823 , #6846 ) @bastisar , @gareth-brown-86 , @mk406
holiday (transformations soft dependency) bounds have been updated to >=0.29,<0.54
dask (data container and parallelization back-end) bounds have been updated to <2024.8.1
implementers no longer need to set the package_import_alias tag when estimator dependencies have a different import name than the PEP 440 package name. All internal logic now only uses the PEP 440 package name. There is no need to remove the tag if already set, but it is no longer required.
estimators now have a tag capability:categorical_in_X: bool to indicate that the estimator can handle categorical features in the input data X . Such estimator can be used with categorical and string-valued features if X is passed in one of the pandas based mtypes.
the html representation of all objects now includes a link to the documentation of the object, and is now in line with the sklearn html representation.
[ENH] improved environment package version check ( #6776 ) @fkiraly
[ENH] Remove package import alias related internal logic and tags ( #6821 ) @fkiraly
[ENH] Adding tag for categorical support in X ( #6704 ) @Abhay-Lejith
[ENH] Adding categorical support: Raising error in yes/no case ( #6732 ) @Abhay-Lejith
[ENH] Link to docs in object’s html repr ( #6876 ) @mateuszkasprowicz
[ENH] Data Loader for M5 dataset ( #6731 ) @SaiRevanth25
[ENH] check_pdmultiindex_panel to return names of invalid object columns if there are any ( #6797 ) @SaiRevanth25
[ENH] Allow object dtype in series ( #5886 ) @yarnabrina
[ENH] converter framework tests in datatypes to cover all types, including those requiring soft dependencies ( #6838 ) @fkiraly
[ENH] add missing feature_kind metadata fields to gluonts based data container checkers ( #6861 ) @fkiraly
[ENH] added feature_kind metadata in datatype checks ( #6490 ) @Abhay-Lejith
[ENH] Adding support for gluonts PandasDataset object ( #6668 ) @shlok191
[ENH] Added support for gluonts PandasDataset as a Series scitype ( #6837 ) @shlok191
[ENH] interface to autots ensemble ( #5948 ) @MBristle
[ENH] darts Reduction Models adapter ( #6712 ) @fnhirwa , @yarnabrina
[ENH] Extension Template For Global Forecasting API ( #6699 ) @XinyuWuu
[ENH] enable multivariate data passed to autots interface ( #6805 ) @fkiraly
[ENH] Add Sampler to ForecastingOptunaSearchCV ( #6823 ) @bastisar
[ENH] Improve TestAllGlobalForecasters ( #6845 ) @XinyuWuu
[ENH] Add scoring direction to ForecastingOptunaSearchCV ( #6846 ) @gareth-brown-86 , @mk406
[ENH] de-novo implementation of LTSFTransformer based on cure-lab research code base ( #6202 ) @geetu040
[ENH] Add windows_identical to DirectReductionForecaster ( #6650 ) @hliebert
[ENH] updates type inference in make_reduction to use central scitype inference and allow proba tabular regressors ( #6893 ) @fkiraly
[ENH] DeepAR and NHiTS and refinements for pytorch-forecasting interface ( #6551 ) @XinyuWuu
[ENH] Interface to TinyTimeMixer foundation model ( #6712 ) @geetu040
[ENH] remove now superfluous try-excepts in forecasting API test suite ( #6906 ) @fkiraly
[ENH] improve test_global_forecasting_tag ( #6929 ) @geetu040
[ENH] in estimator html repr, make version retrieval safer and more flexible ( #6923 ) @fkiraly
[ENH] time series annotation (outliers, changepoints) - test class and full check_estimator integration ( #6843 ) @fkiraly
[ENH] Add Windowed Local Outlier Factor Anomaly Detector ( #6524 ) @Alex-JG3
[ENH] Add binary segmentation annotator for change point detection ( #6723 ) @Alex-JG3
[ENH] Pytorch Classifier intermediate base class for TSC ( #6791 ) @geetu040
[ENH] MVTS transformer classifier ( #6791 ) @geetu040
[ENH] add second test params dict to Aggregator ( #6759 ) @fr1ll
[ENH] pandas inner type and global pooling for TabularToSeriesAdaptor ( #6752 ) @fkiraly
[ENH] alternative returns for VmdTransformer - mode spectra and central frequencies ( #6857 ) @fkiraly
[ENH] simplify dictionaries and alias handling in Catch22 ( #6104 ) @fkiraly
[ENH] making self._is_vectorized access more defensive in BaseTransformer ( #6863 ) @fkiraly
[ENH] make pyproject.toml parsing for differential testing more robust against non-package relevant changes ( #6882 ) @fkiraly
[ENH] Vendor fracdiff library ( #6777 ) @DinoBektesevic
[ENH] improvements to vendored fracdiff library ( #6912 ) @fkiraly
[DOC] Notebook and Template For Global Forecasting API ( #6699 ) @XinyuWuu
[DOC] Add authorship credits to MatrixProfileTransformer for Stumpy authors ( #6762 ) @alexander-lakocy
[DOC] add examples to StatsForecastGARCH and StatsForecastARCH docstrings ( #6761 ) @melinny
[DOC] Add alignment notebook example ( #6768 ) @alexander-lakocy
[DOC] fix transformers type table in API reference in accordance with sphinx guidelines ( #6771 ) @alexander-lakocy
[DOC] Modify editable install to make cross-platform ( #6758 ) @fr1ll
[DOC] TruncationTransformer docstring example ( #6765 ) @ceroper
[DOC] De-duplicate User Guide and Examples (closes #6767) ( #6770 ) @alexander-lakocy
[DOC] improved docstring of DWTTransformer ( #6764 ) @Mitchjkjkjk
[DOC] various improvements to user journey on documentation page ( #6760 ) @fkiraly
[DOC] Time series k means max iter parameter docstring ( #6726 ) @AlexeyOm
[DOC] cross-reference estimator search from tags API reference ( #6816 ) @fkiraly , @yarnabrina
[DOC] updated docstring for check_is_mtype to match skpro check_is_mtype function ( #6835 ) @julian-fong
[DOC] example & tutorial notebooks: normalize execution counts, indentation, execute all cells ( #6847 ) @fkiraly
[DOC] clarify column handling in docstring of FourierFeatures ( #6834 ) @fkiraly
[DOC] added fork usage recommendations ( #6827 ) @yarnabrina
[DOC] change links in documentation to refer to same version ( #6841 ) @yarnabrina
[DOC] minor improvements to check_scoring docstring ( #6877 ) @fkiraly
[DOC] add proper author credits to 1:1 interface classes - aligners, distances, forecasters, parameter estimators ( #6850 ) @fkiraly
[DOC] fix docstring formatting of evaluate ( #6864 ) @fkiraly
[DOC] Add documentation for benchmarking module ( #6792 ) @benHeid
[DOC] add elections link on landing page ( #6910 ) @fkiraly
[DOC] Add example notebook for the graphical pipeline ( #5175 ) @benHeid
[DOC] git workflow guide - chained branches, fixing header fonts ( #6913 ) @fkiraly
[MNT] Remove tbats python version constraint ( #6769 ) @fr1ll
[MNT] Update Callable import from typing to collections.abc ( #6798 ) @yarnabrina
[MNT] Fix spellings using codespell and typos ( #6799 ) @yarnabrina
[MNT] improved environment package version check ( #6776 ) @fkiraly
[MNT] downgrade pykan version to <0.2.2 ( #6853 ) @geetu040
[MNT] add non-unicode characters check to the linter ( #6807 ) @fnhirwa
[MNT] updates and fixes to type hints ( #6743 ) @ZhipengXue97
[MNT] Resolve the issue with diacritics failing to be decoded on Windows ( #6862 ) @fnhirwa
[MNT] sync docstring and code formatting of dependency checker module with skbase ( #6873 ) @fkiraly
[MNT] Remove package import alias related internal logic and tags ( #6821 ) @fkiraly
[MNT] restrict failing Mr-SEQL version ( #6879 ) @fkiraly
[MNT] release workflow: Upgrade deprecated pypa action parameter ( #6878 ) @szepeviktor
[MNT] Fix pykan import and dependency checks ( #6881 ) @fkiraly
[MNT] temporarily pin matplotlib below 3.9.1 ( #6890 ) @yarnabrina
[MNT] make pyproject.toml parsing for differential testing more robust against non-package relevant changes ( #6882 ) @fkiraly
[MNT] formatter for jupyter notebook json in build tools ( #6849 ) @fkiraly
[MNT] sync differential testing utilities with skpro ( #6840 ) @fkiraly
[MNT] Handle deprecations from pandas ( #6855 ) @fkiraly
[MNT] sync docstring and code formatting of dependency checker module with skbase ( #6873 ) @fkiraly
[MNT] fix .all-contributorsrc syntax ( #6918 ) @fkiraly
[MNT] Resolve the issue with diacritics failing to be decoded on Windows ( #6862 ) @fnhirwa
[MNT] changelog utility: fix termination condition to retrieve merged PR ( #6920 ) @fkiraly
[MNT] restore holidays lower bound to 0.29 ( #6921 ) @fkiraly
[MNT] Updating the GHA dependencies to install OSX dependencies and setting the compiler flags ( #6926 ) @fnhirwa
[MNT] revert an erroneous instance of pandas deprecation fix ( #6925 ) @fkiraly
[MNT] Update the path to script to fix #6926 ( #6933 ) @fnhirwa
[MNT] Dependabot: Update pytest requirement from <8.3,>=7.4 to >=7.4,<8.4 ( #6819 ) @dependabot[bot]
[MNT] Dependabot: Update dask requirement from <2024.6.3 to <2024.7.2 ( #6818 ) @dependabot[bot]
[MNT] Dependabot: Update sphinx-gallery requirement from <0.17.0 to <0.18.0 ( #6820 ) @dependabot[bot]
[MNT] Dependabot: Update holidays requirement from <0.53,>=0.52 to >=0.52,<0.54 ( #6780 ) @dependabot[bot]
[MNT] Dependabot: Update sphinx requirement from !=7.2.0,<8.0.0 to !=7.2.0,<9.0.0 ( #6865 ) @dependabot[bot]
[MNT] Dependabot: Update holidays requirement from <0.54,>=0.52 to >=0.52,<0.55 ( #6898 ) @dependabot[bot]
[MNT] Dependabot: Update dask requirement from <2024.7.2 to <2024.8.1 ( #6907 ) @dependabot[bot]
[BUG] fix _check_soft_dependencies for post and pre versions of patch versions ( #6909 ) @fkiraly
[BUG] fix type inconsistency in conversion pandas to xarray based Series ( #6856 ) @fkiraly
[BUG] Fix pykan dependency and set lower bound ( #6789 ) @benHeid
[BUG] correct dependency tag for pytorch-forecasting forecasters: rename pytorch_forecasting to correct package name pytorch-forecasting ( #6830 ) @XinyuWuu
[BUG] fix polymorphic estimators missing in estimator overview, e.g., pytorch-forecasting forecasters ( #6803 ) @fkiraly
[BUG] Fix bug when predicting segments from clasp change point annotator ( #6756 ) @Alex-JG3
[BUG] Refactor ADICVTransformer and fix CV calculation ( #6757 ) @sbhobbes
[BUG] fix BaseTransformer broadcasting condition in inverse_transform for decomposers ( #6824 ) @fkiraly
[BUG] fix MSTL inverse transform and use in forecasting pipeline ( #6825 ) @fkiraly
[BUG] fix handling of numpy integers in refactored Catch22 transformation ( #6934 ) @fkiraly
[BUG] In plot_series , trim unused levels when verifying dataframe formatting ( #6754 ) @SultanOrazbayev
@Abhay-Lejith , @Alex-JG3 , @alexander-lakocy , @AlexeyOm , @bastisar , @benHeid , @ceroper , @DinoBektesevic , @fkiraly , @fnhirwa , @fr1ll , @gareth-brown-86 , @geetu040 , @hliebert , @julian-fong , @mateuszkasprowicz , @MBristle , @melinny , @Mitchjkjkjk , @mk406 , @SaiRevanth25 , @sbhobbes , @shlok191 , @SultanOrazbayev , @szepeviktor , @XinyuWuu , @yarnabrina , @ZhipengXue97
html representation of objects now has a button linking to documentation page (6876) mateuszkasprowicz
interface to TinyTimeMixer foundation model (6712) geetu040
interface to autots ensemble (5948) MBristle
interface to darts reduction models (6712) fnhirwa, yarnabrina
LTSFTransformer based on cure-lab research code base (6202) geetu040
MVTS transformer classifier (6791) geetu040
forecasters can now support categorical X, as per tag (6704, 6732) Abhay-Lejith
DirectReductionForecaster now has a windows_identical option (6650) hliebert
ForecastingOptunaSearchCV can now be passed custom samplers and "higher is better" scores (6823, 6846) bastisar, gareth-brown-86, mk406
holiday (transformations soft dependency) bounds have been updated to >=0.29,<0.54
dask (data container and parallelization back-end) bounds have been updated to <2024.8.1
implementers no longer need to set the package_import_alias tag when estimator dependencies have a different import name than the PEP 440 package name. All internal logic now only uses the PEP 440 package name. There is no need to remove the tag if already set, but it is no longer required.
estimators now have a tag capability:categorical_in_X: bool to indicate that the estimator can handle categorical features in the input data X. Such estimator can be used with categorical and string-valued features if X is passed in one of the pandas based mtypes.
the html representation of all objects now includes a link to the documentation of the object, and is now in line with the sklearn html representation.
[ENH] improved environment package version check (6776) fkiraly
[ENH] Remove package import alias related internal logic and tags (6821) fkiraly
[ENH] Adding tag for categorical support in X (6704) Abhay-Lejith
[ENH] Adding categorical support: Raising error in yes/no case (6732) Abhay-Lejith
[ENH] Link to docs in object's html repr (6876) mateuszkasprowicz
[ENH] Data Loader for M5 dataset (6731) SaiRevanth25
[ENH] check_pdmultiindex_panel to return names of invalid object columns if there are any (6797) SaiRevanth25
[ENH] Allow object dtype in series (5886) yarnabrina
[ENH] converter framework tests in datatypes to cover all types, including those requiring soft dependencies (6838) fkiraly
[ENH] add missing feature_kind metadata fields to gluonts based data container checkers (6861) fkiraly
[ENH] added feature_kind metadata in datatype checks (6490) Abhay-Lejith
[ENH] Adding support for gluonts PandasDataset object (6668) shlok191
[ENH] Added support for gluonts PandasDataset as a Series scitype (6837) shlok191
[ENH] interface to autots ensemble (5948) MBristle
[ENH] darts Reduction Models adapter (6712) fnhirwa, yarnabrina
[ENH] Extension Template For Global Forecasting API (6699) XinyuWuu
[ENH] enable multivariate data passed to autots interface (6805) fkiraly
[ENH] Add Sampler to ForecastingOptunaSearchCV (6823) bastisar
[ENH] Improve TestAllGlobalForecasters (6845) XinyuWuu
[ENH] Add scoring direction to ForecastingOptunaSearchCV (6846) gareth-brown-86, mk406
[ENH] de-novo implementation of LTSFTransformer based on cure-lab research code base (6202) geetu040
[ENH] Add windows_identical to DirectReductionForecaster (6650) hliebert
[ENH] updates type inference in make_reduction to use central scitype inference and allow proba tabular regressors (6893) fkiraly
[ENH] DeepAR and NHiTS and refinements for pytorch-forecasting interface (6551) XinyuWuu
[ENH] Interface to TinyTimeMixer foundation model (6712) geetu040
[ENH] remove now superfluous try-excepts in forecasting API test suite (6906) fkiraly
[ENH] improve test_global_forecasting_tag (6929) geetu040
[ENH] in estimator html repr, make version retrieval safer and more flexible (6923) fkiraly
[ENH] time series annotation (outliers, changepoints) - test class and full check_estimator integration (6843) fkiraly
[ENH] Add Windowed Local Outlier Factor Anomaly Detector (6524) Alex-JG3
[ENH] Add binary segmentation annotator for change point detection (6723) Alex-JG3
[ENH] Pytorch Classifier intermediate base class for TSC (6791) geetu040
[ENH] MVTS transformer classifier (6791) geetu040
[ENH] add second test params dict to Aggregator (6759) fr1ll
[ENH] pandas inner type and global pooling for TabularToSeriesAdaptor (6752) fkiraly
[ENH] alternative returns for VmdTransformer - mode spectra and central frequencies (6857) fkiraly
[ENH] simplify dictionaries and alias handling in Catch22 (6104) fkiraly
[ENH] making self._is_vectorized access more defensive in BaseTransformer (6863) fkiraly
[ENH] make pyproject.toml parsing for differential testing more robust against non-package relevant changes (6882) fkiraly
[ENH] Vendor fracdiff library (6777) DinoBektesevic
[ENH] improvements to vendored fracdiff library (6912) fkiraly
[DOC] Notebook and Template For Global Forecasting API (6699) XinyuWuu
[DOC] Add authorship credits to MatrixProfileTransformer for Stumpy authors (6762) alexander-lakocy
[DOC] add examples to StatsForecastGARCH and StatsForecastARCH docstrings (6761) melinny
[DOC] Add alignment notebook example (6768) alexander-lakocy
[DOC] fix transformers type table in API reference in accordance with sphinx guidelines (6771) alexander-lakocy
[DOC] Modify editable install to make cross-platform (6758) fr1ll
[DOC] TruncationTransformer docstring example (6765) ceroper
[DOC] De-duplicate User Guide and Examples (closes #6767) (6770) alexander-lakocy
[DOC] improved docstring of DWTTransformer (6764) Mitchjkjkjk
[DOC] various improvements to user journey on documentation page (6760) fkiraly
[DOC] Time series k means max iter parameter docstring (6726) AlexeyOm
[DOC] cross-reference estimator search from tags API reference (6816) fkiraly, yarnabrina
[DOC] updated docstring for check_is_mtype to match skpro check_is_mtype function (6835) julian-fong
[DOC] example & tutorial notebooks: normalize execution counts, indentation, execute all cells (6847) fkiraly
[DOC] clarify column handling in docstring of FourierFeatures (6834) fkiraly
[DOC] added fork usage recommendations (6827) yarnabrina
[DOC] change links in documentation to refer to same version (6841) yarnabrina
[DOC] minor improvements to check_scoring docstring (6877) fkiraly
[DOC] add proper author credits to 1:1 interface classes - aligners, distances, forecasters, parameter estimators (6850) fkiraly
[DOC] fix docstring formatting of evaluate (6864) fkiraly
[DOC] Add documentation for benchmarking module (6792) benHeid
[DOC] add elections link on landing page (6910) fkiraly
[DOC] Add example notebook for the graphical pipeline (5175) benHeid
[DOC] git workflow guide - chained branches, fixing header fonts (6913) fkiraly
[MNT] Remove tbats python version constraint (6769) fr1ll
[MNT] Update Callable import from typing to collections.abc (6798) yarnabrina
[MNT] Fix spellings using codespell and typos (6799) yarnabrina
[MNT] improved environment package version check (6776) fkiraly
[MNT] downgrade pykan version to <0.2.2 (6853) geetu040
[MNT] add non-unicode characters check to the linter (6807) fnhirwa
[MNT] updates and fixes to type hints (6743) ZhipengXue97
[MNT] Resolve the issue with diacritics failing to be decoded on Windows (6862) fnhirwa
[MNT] sync docstring and code formatting of dependency checker module with skbase (6873) fkiraly
[MNT] Remove package import alias related internal logic and tags (6821) fkiraly
[MNT] restrict failing Mr-SEQL version (6879) fkiraly
[MNT] release workflow: Upgrade deprecated pypa action parameter (6878) szepeviktor
[MNT] Fix pykan import and dependency checks (6881) fkiraly
[MNT] temporarily pin matplotlib below 3.9.1 (6890) yarnabrina
[MNT] make pyproject.toml parsing for differential testing more robust against non-package relevant changes (6882) fkiraly
[MNT] formatter for jupyter notebook json in build tools (6849) fkiraly
[MNT] sync differential testing utilities with skpro (6840) fkiraly
[MNT] Handle deprecations from pandas (6855) fkiraly
[MNT] sync docstring and code formatting of dependency checker module with skbase (6873) fkiraly
[MNT] fix .all-contributorsrc syntax (6918) fkiraly
[MNT] Resolve the issue with diacritics failing to be decoded on Windows (6862) fnhirwa
[MNT] changelog utility: fix termination condition to retrieve merged PR (6920) fkiraly
[MNT] restore holidays lower bound to 0.29 (6921) fkiraly
[MNT] Updating the GHA dependencies to install OSX dependencies and setting the compiler flags (6926) fnhirwa
[MNT] revert an erroneous instance of pandas deprecation fix (6925) fkiraly
[MNT] Update the path to script to fix #6926 (6933) fnhirwa
[MNT] [Dependabot](deps): Update pytest requirement from <8.3,>=7.4 to >=7.4,<8.4 (6819) dependabot[bot]
[MNT] [Dependabot](deps): Update dask requirement from <2024.6.3 to <2024.7.2 (6818) dependabot[bot]
[MNT] [Dependabot](deps): Update sphinx-gallery requirement from <0.17.0 to <0.18.0 (6820) dependabot[bot]
[MNT] [Dependabot](deps): Update holidays requirement from <0.53,>=0.52 to >=0.52,<0.54 (6780) dependabot[bot]
[MNT] [Dependabot](deps): Update sphinx requirement from !=7.2.0,<8.0.0 to !=7.2.0,<9.0.0 (6865) dependabot[bot]
[MNT] [Dependabot](deps): Update holidays requirement from <0.54,>=0.52 to >=0.52,<0.55 (6898) dependabot[bot]
[MNT] [Dependabot](deps): Update dask requirement from <2024.7.2 to <2024.8.1 (6907) dependabot[bot]
[BUG] fix _check_soft_dependencies for post and pre versions of patch versions (6909) fkiraly
[BUG] fix type inconsistency in conversion pandas to xarray based Series (6856) fkiraly
[BUG] Fix pykan dependency and set lower bound (6789) benHeid
[BUG] correct dependency tag for pytorch-forecasting forecasters: rename pytorch_forecasting to correct package name pytorch-forecasting (6830) XinyuWuu
[BUG] fix polymorphic estimators missing in estimator overview, e.g., pytorch-forecasting forecasters (6803) fkiraly
[BUG] Fix bug when predicting segments from clasp change point annotator (6756) Alex-JG3
[BUG] Refactor ADICVTransformer and fix CV calculation (6757) sbhobbes
[BUG] fix BaseTransformer broadcasting condition in inverse_transform for decomposers (6824) fkiraly
[BUG] fix MSTL inverse transform and use in forecasting pipeline (6825) fkiraly
[BUG] fix handling of numpy integers in refactored Catch22 transformation (6934) fkiraly
[BUG] In plot_series, trim unused levels when verifying dataframe formatting (6754) SultanOrazbayev
Abhay-Lejith, Alex-JG3, alexander-lakocy, AlexeyOm, bastisar, benHeid, ceroper, DinoBektesevic, fkiraly, fnhirwa, fr1ll, gareth-brown-86, geetu040, hliebert, julian-fong, mateuszkasprowicz, MBristle, melinny, Mitchjkjkjk, mk406, SaiRevanth25, sbhobbes, shlok191, SultanOrazbayev, szepeviktor, XinyuWuu, yarnabrina, ZhipengXue97
scheduled deprecations and change actions
Maintenance release with numpy 2 compatibility of framework layer.
ruff and python >= 3.9 based precommitsFor last larger feature update, see 0.30.2.
Please see our changelog for a description of all changes.
@fkiraly, @fnhirwa, @yarnabrina
Full Changelog: https://github.com/sktime/sktime/compare/v0.30.2...v0.31.0
Maintenance release:
scheduled deprecations and change actions
numpy 2 compatibility
code style and pre-commit updates, using ruff for linting
For last non-maintenance content updates, see 0.30.2.
numpy (core dependency) bounds have been updated to <2.1,>=1.21
skpro (soft dependency) bounds have been updated to >=2,<2.5.0
The fmt argument in time series annotators is now deprecated. Users should use the predict and transform methods instead, predict instead of fmt="sparse" , and transform instead of fmt="dense" .
The convert_y_to_keras method in deep learning classifiers has been removed. Users who have been using this method should instead use OneHotEncoder from sklearn directly, as convert_y_to_keras is a simple wrapper around OneHotEncoder with default settings.
[MNT] raise numpy bound to numpy < 2.1 , numpy 2 compatibility ( #6624 ) @fkiraly
[MNT] Dependabot: Update skpro requirement from <2.4.0,>=2 to >=2,<2.5.0 ( #6663 ) @dependabot[bot]
[MNT] bound prophet based forecasters to numpy<2 due to incompatibility of prophet ( #6721 ) @fkiraly
[MNT] further numpy 2 compatibility fixes in estimators ( #6729 ) @fkiraly
[MNT] handle numpy 2 incompatible soft deps ( #6728 ) @fkiraly
[MNT] Upgrade code style beyond python 3.8 ( #6330 ) @yarnabrina
[MNT] Update pre commit hooks post dropping python 3.8 support ( #6331 ) @yarnabrina
[MNT] suppress aggressive freq related warnings from pandas 2.2 ( #6733 ) @fkiraly
[MNT] 0.31.0 deprecations and change actions ( #6716 ) @fkiraly
[MNT] switch to ruff as linting tool ( #6676 ) @fnhirwa
[ENH] refactor and bugfixes for environment checker utilities ( #6719 ) @fkiraly
@fkiraly , @fnhirwa , @yarnabrina
Maintenance release:
scheduled deprecations and change actions
numpy 2 compatibility
code style and pre-commit updates, using ruff for linting
For last non-maintenance content updates, see 0.30.2.
numpy (core dependency) bounds have been updated to <2.1,>=1.21
skpro (soft dependency) bounds have been updated to >=2,<2.5.0
The fmt argument in time series annotators is now deprecated. Users should use the predict and transform methods instead, predict instead of fmt="sparse", and transform instead of fmt="dense".
The convert_y_to_keras method in deep learning classifiers has been removed. Users who have been using this method should instead use OneHotEncoder from sklearn directly, as convert_y_to_keras is a simple wrapper around OneHotEncoder with default settings.
[MNT] raise numpy bound to numpy < 2.1, numpy 2 compatibility (6624) fkiraly
[MNT] [Dependabot](deps): Update skpro requirement from <2.4.0,>=2 to >=2,<2.5.0 (6663) dependabot[bot]
[MNT] bound prophet based forecasters to numpy<2 due to incompatibility of prophet (6721) fkiraly
[MNT] further numpy 2 compatibility fixes in estimators (6729) fkiraly
[MNT] handle numpy 2 incompatible soft deps (6728) fkiraly
[MNT] Upgrade code style beyond python 3.8 (6330) yarnabrina
[MNT] Update pre commit hooks post dropping python 3.8 support (6331) yarnabrina
[MNT] suppress aggressive freq related warnings from pandas 2.2 (6733) fkiraly
[MNT] 0.31.0 deprecations and change actions (6716) fkiraly
[MNT] switch to ruff as linting tool (6676) fnhirwa
[ENH] refactor and bugfixes for environment checker utilities (6719) fkiraly
fkiraly, fnhirwa, yarnabrina
Please see our changelog for a description of all changes.
Please see our changelog for a description of all changes.
@Abhay-Lejith, @benHeid, @benshaw2, @doberbauer, @emmanuel-ferdman, @ericjb, @felipeangelimvieira, @fkiraly, @fnhirwa, @gareth-brown-86, @geetu040, @iaryangoyal, @julian-fong, @ksharma6, @mk406, @shlok191, @Spinachboul, @TheoWeih, @Xinyu-Wu-0000, @yarnabrina, @Z-Fran
Full Changelog: https://github.com/sktime/sktime/compare/v0.30.1...v0.30.2
new estimator overview table and estimator search page ( #6147 ) @duydl
HFTransformersForecaster (hugging face transformers connector) now has a user friendly interface for applying PEFT methods ( #6457 ) @geetu040
ForecastingOptunaSearchCV for hyper-parameter tuning of forecasters via optuna ( #6630 ) @mk406 , @gareth-brown-86
prophetverse package forecasters are now indexed by sktime ( #6614 ) @felipeangelimvieira
pytorch-forecasting adapter, experimental global forecasting API ( #6228 ) @XinyuWuu
skforecast adapter for reduction strategies ( #6531 ) @Abhay-Lejith , @yarnabrina
EnbPI based forecaster with components from aws-fortuna ( #6449 ) @benHeid
DTW distances and aligners from dtaidistance ( #6578 ) @fkiraly
parametrize_with_checks utility for granular API compliance test setup in 2nd/3rd party libraries ( #6588 ) @fkiraly
holidays (transformations soft dependency) bounds have been updated to >=0.29,<0.53
dask (data container and parallelization back-end) bounds have been updated to <2024.5.3
optuna is now a soft dependency, via the ForecastingOptunaSearchCV estimator, in the all_extras soft dependency set, with bounds <3.7
pytorch-forecasting is now a soft dependency, in the dl (deep learning) soft dependency set
skforecast is now a soft dependency, in the all_extras soft dependency set and the forecasting soft dependency set, with bounds <0.13,>=0.12.1
dtaidistance is now a soft dependency, in the all_extras soft dependency set and the alignment soft dependency set, with bounds <2.4
The base forecaster interface now has a dedicated interface point for global forecasting or fine-tuning: in forecasters supporting global forecast, an y argument may be passed in predict , indicating new time series instances for a global forecast, or a context for foundation models. Forecasters capable of global forecasting or fine-tuning (this is the same interface point) are tagged with the tag capability:global_forecasting , value True .
The global forecasting and fine-tuning interfaces are currently experimental, and may undergo changes.
Users are invited to give feedback, and test the feature with the new pytorch-forecasting adapter.
2nd and 3rd party extension packages can now use the parametrize_with_checks utility to set up granular API compliance tests. For detailed usage notes, consult the extender guide: Implementing Estimators .
various quality-of-life improvements have been made to facilitate indexing an estimator in the estimator overview and estimator search for developers of API compatible 2nd and 3rd party packages, without adding it directly to the main sktime repository. For detailed usage notes, consult the extender guide: Implementing Estimators , or inspect the Prophetverse forecaster as a worked example.
[ENH] prevent imports caused by _check_soft_dependencies , speed up dependency check and test collection time ( #6355 ) @fkiraly , @yarnabrina
[ENH] Parallelization option for ForecastingBenchmark ( #6568 ) @benHeid
[ENH] Added GluonTS datasets as sktime mtypes ( #6530 ) @shlok191
[ENH] DTW distances from dtaidistance ( #6578 ) @fkiraly
[ENH] pytorch-forecasting adapter with Global Forecasting API ( #6228 ) @XinyuWuu
[ENH] fitted parameter forwarding utility, forward statsforecast estimators’ fitted parameters ( #6349 ) @fkiraly
[ENH] EnbPI based forecaster with components from aws-fortuna ( #6449 ) @benHeid
[ENH] skforecast ForecasterAutoreg adapter ( #6531 ) @Abhay-Lejith , @yarnabrina
[ENH] Extend HFTransformersForecaster for PEFT methods ( #6457 ) @geetu040
[ENH] in BaseForecaster , move check for capability:insample to _check_fh boilerplate ( #6593 ) @XinyuWuu
[ENH] indexing prophetverse forecaster ( #6614 ) @fkiraly
[ENH] ForecastingOptunaSearchCV for hyper-parameter tuning of forecasters via optuna ( #6630 ) @mk406 , @gareth-brown-86
[ENH] enhanced estimator overview table - tag display and search ( #6147 ) @duydl , @fkiraly
[ENH] DTW aligners from dtaidistance ( #6578 ) @fkiraly
[ENH] resolve duplication in KNeighborsClassifier and KNeighborsRegressor ( #6504 ) @Z-Fran
[ENH] added two test params sets to FCNNetwork ( #6562 ) @TheoWeih
[ENH] further refactor of knn classifier and regressor ( #6615 ) @fkiraly
[ENH] update tests._config to skip various sporadically failing tests for Proximity Forest and Proximity Tree until fixed ( #6638 ) @julian-fong
[ENH] Time Series Regression grid search ( #6118 ) @ksharma6
[ENH] test parameters for RocketRegressor ( #6149 ) @iaryangoyal
[ENH] resolve duplication in KNeighborsClassifier and KNeighborsRegressor ( #6504 ) @Z-Fran
[ENH] further refactor of knn classifier and regressor ( #6615 ) @fkiraly
[ENH] refactor WindowSummarizer tests ( #6564 ) @fkiraly
[ENH] differential testing for base functionality in various modules ( #6534 ) @fkiraly
[ENH] further differential testing for the transformations module ( #6533 ) @fkiraly
[ENH] differential testing in dist_kernels and clustering modules ( #6543 ) @fkiraly
[ENH] simplify and add differential testing to forecasting.compose.tests module ( #6563 ) @fkiraly
[ENH] simplify and add differential testing to sktime.pipeline module ( #6565 ) @fkiraly
[ENH] differential testing in benchmarking module ( #6566 ) @fkiraly
[ENH] move doctests to main test suite to ensure conditional execution ( #6536 ) @fkiraly
[ENH] minor improvements to test efficiency ( #6586 ) @fkiraly
[ENH] parametrize_with_checks utility for granular API compliance test setup in 2nd/3rd party libraries ( #6588 ) @fkiraly
[ENH] differential testing to utils module ( #6620 ) @fkiraly
[ENH] differential testing and minor improvements to forecasting.base tests ( #6619 ) @fkiraly
[ENH] differential testing for performance_metrics module ( #6616 ) @fkiraly
[ENH] fixes and improvements to pytest doctest integration ( #6621 ) @fkiraly
[DOC] fix broken links on webpage docs ( #6339 ) @duydl
[DOC] document more tags ( #6496 ) @fkiraly
[DOC] fix minor typos in tags API reference ( #6631 ) @fkiraly
[DOC] update dependencies reference ( #6655 ) @emmanuel-ferdman
[DOC] fix minor typo in developer comment in BaseTransformer ( #6689 ) @Spinachboul
[DOC] rst roadmap documentation page stale since 2021 - replace by correct links to recent roadmaps ( #6556 ) @fkiraly
[DOC] clarify docs on ARIMA estimators, add author credits for upstream ( #6705 ) @fkiraly
[DOC] credit @doberbauer` for pykalman python 3.11 compatibility fix ( #6662 ) @doberbauer
[MNT] Dependabot: Update holidays requirement from <0.51,>=0.29 to >=0.29,<0.52 ( #6634 ) @dependabot[bot]
[MNT] Dependabot: Update holidays requirement from <0.52,>=0.29 to >=0.52,<0.53 ( #6702 ) @dependabot[bot]
[MNT] Dependabot: Update dask requirement from <2024.6.1 to <2024.6.2 ( #6643 ) @dependabot[bot]
[MNT] Dependabot: Update numba requirement from <0.60,>=0.53 to >=0.53,<0.61 ( #6590 ) @dependabot[bot]
[MNT] Dependabot: Update dask requirement from <2024.6.2 to <2024.6.3 ( #6647 ) @dependabot[bot]
[MNT] remove coverage reporting and pytest-cov from PR CI and setup.cfg ( #6363 ) @fkiraly
[MNT] numpy 2 compatibility fixes - estimators ( #6626 ) @fkiraly
[MNT] scipy 1.14.0 compatibility for deep_equals plugin for csr_matrix ( #6664 ) @fkiraly
[MNT] deprecate unused _check_soft_dependencies argument suppress_import_stdout ( #6691 ) @fkiraly
[BUG] fix AUCalibration probabilistic metric for multivariate case ( #6617 ) @fkiraly
[BUG] fix bug 4076: PerformanceWarning in load_from_tsfile_to_dataframe ( #6632 ) @ericjb
[BUG] patch over pandas 2.2.X issue in freq timestamp/period round trip conversion for period start timestamps such as "MonthBegin" ( #6574 ) @fkiraly
[BUG] fix passing of y in ForecastingPipeline ( #6706 ) @fkiraly
[BUG] Fix bug in fitted parameter override in pyts and tslearn adapters ( #6707 ) @fkiraly
[BUG] Fix bug in fitted parameter override in pyts and tslearn adapters ( #6707 ) @fkiraly
[BUG] in TimeSeriesForestRegressor , fix failure: self.criterion does not exist ( #6573 ) @ksharma6
[BUG] partially revert pytest.skip change from #6233 due to side effects in downstream test suites ( #6508 ) @fkiraly
[BUG] fix test failures introduced by differential testing refactor ( #6585 ) @fkiraly
[BUG] fix HolidayFeatures crashes if dataframe doesn’t contain specified date ( #6550 ) @fnhirwa
[BUG] in Differencer , make explicit clone to avoid SettingWithCopyWarning ( #6567 ) @benHeid
[BUG] minirocket: fix zero division errors #5174 ( #6612 ) @benshaw2
[BUG] ensure correct setting of requires_X and requires_y tag for FeatureUnion ( #6695 ) @fkiraly
[BUG] ensure correct setting of requires_X and requires_y tag for TransformerPipeline ( #6692 ) @fkiraly
[BUG] partial fix for dropped column names in PaddingTransformer ( #6693 ) @fkiraly
@Abhay-Lejith , @benHeid , @benshaw2 , @doberbauer , @emmanuel-ferdman , @ericjb , @felipeangelimvieira , @fkiraly , @fnhirwa , @gareth-brown-86 , @geetu040 , @iaryangoyal , @julian-fong , @ksharma6 , @mk406 , @shlok191 , @Spinachboul , @TheoWeih , @XinyuWuu , @yarnabrina , @Z-Fran
new estimator overview table and estimator search page (6147) duydl
HFTransformersForecaster (hugging face transformers connector) now has a user friendly interface for applying PEFT methods (6457) geetu040
ForecastingOptunaSearchCV for hyper-parameter tuning of forecasters via optuna (6630) mk406, gareth-brown-86
prophetverse package forecasters are now indexed by sktime (6614) felipeangelimvieira
pytorch-forecasting adapter, experimental global forecasting API (6228) XinyuWuu
skforecast adapter for reduction strategies (6531) Abhay-Lejith, yarnabrina
EnbPI based forecaster with components from aws-fortuna (6449) benHeid
DTW distances and aligners from dtaidistance (6578) fkiraly
parametrize_with_checks utility for granular API compliance test setup in 2nd/3rd party libraries (6588) fkiraly
holidays (transformations soft dependency) bounds have been updated to >=0.29,<0.53
dask (data container and parallelization back-end) bounds have been updated to <2024.5.3
optuna is now a soft dependency, via the ForecastingOptunaSearchCV estimator, in the all_extras soft dependency set, with bounds <3.7
pytorch-forecasting is now a soft dependency, in the dl (deep learning) soft dependency set
skforecast is now a soft dependency, in the all_extras soft dependency set and the forecasting soft dependency set, with bounds <0.13,>=0.12.1
dtaidistance is now a soft dependency, in the all_extras soft dependency set and the alignment soft dependency set, with bounds <2.4
The base forecaster interface now has a dedicated interface point for global forecasting or fine-tuning: in forecasters supporting global forecast, an y argument may be passed in predict, indicating new time series instances for a global forecast, or a context for foundation models. Forecasters capable of global forecasting or fine-tuning (this is the same interface point) are tagged with the tag capability:global_forecasting, value True.
The global forecasting and fine-tuning interfaces are currently experimental, and may undergo changes.
Users are invited to give feedback, and test the feature with the new pytorch-forecasting adapter.
2nd and 3rd party extension packages can now use the parametrize_with_checks utility to set up granular API compliance tests. For detailed usage notes, consult the extender guide: developer_guide_add_estimators.
various quality-of-life improvements have been made to facilitate indexing an estimator in the estimator overview and estimator search for developers of API compatible 2nd and 3rd party packages, without adding it directly to the main sktime repository. For detailed usage notes, consult the extender guide: developer_guide_add_estimators, or inspect the Prophetverse forecaster as a worked example.
[ENH] prevent imports caused by _check_soft_dependencies, speed up dependency check and test collection time (6355) fkiraly, yarnabrina
[ENH] Parallelization option for ForecastingBenchmark (6568) benHeid
[ENH] Added GluonTS datasets as sktime mtypes (6530) shlok191
[ENH] DTW distances from dtaidistance (6578) fkiraly
[ENH] pytorch-forecasting adapter with Global Forecasting API (6228) XinyuWuu
[ENH] fitted parameter forwarding utility, forward statsforecast estimators' fitted parameters (6349) fkiraly
[ENH] EnbPI based forecaster with components from aws-fortuna (6449) benHeid
[ENH] skforecast ForecasterAutoreg adapter (6531) Abhay-Lejith, yarnabrina
[ENH] Extend HFTransformersForecaster for PEFT methods (6457) geetu040
[ENH] in BaseForecaster, move check for capability:insample to _check_fh boilerplate (6593) XinyuWuu
[ENH] indexing prophetverse forecaster (6614) fkiraly
[ENH] ForecastingOptunaSearchCV for hyper-parameter tuning of forecasters via optuna (6630) mk406, gareth-brown-86
[ENH] enhanced estimator overview table - tag display and search (6147) duydl, fkiraly
[ENH] DTW aligners from dtaidistance (6578) fkiraly
[ENH] resolve duplication in KNeighborsClassifier and KNeighborsRegressor (6504) Z-Fran
[ENH] added two test params sets to FCNNetwork (6562) TheoWeih
[ENH] further refactor of knn classifier and regressor (6615) fkiraly
[ENH] update tests._config to skip various sporadically failing tests for Proximity Forest and Proximity Tree until fixed (6638) julian-fong
[ENH] Time Series Regression grid search (6118) ksharma6
[ENH] test parameters for RocketRegressor (6149) iaryangoyal
[ENH] resolve duplication in KNeighborsClassifier and KNeighborsRegressor (6504) Z-Fran
[ENH] further refactor of knn classifier and regressor (6615) fkiraly
[ENH] refactor WindowSummarizer tests (6564) fkiraly
[ENH] differential testing for base functionality in various modules (6534) fkiraly
[ENH] further differential testing for the transformations module (6533) fkiraly
[ENH] differential testing in dist_kernels and clustering modules (6543) fkiraly
[ENH] simplify and add differential testing to forecasting.compose.tests module (6563) fkiraly
[ENH] simplify and add differential testing to sktime.pipeline module (6565) fkiraly
[ENH] differential testing in benchmarking module (6566) fkiraly
[ENH] move doctests to main test suite to ensure conditional execution (6536) fkiraly
[ENH] minor improvements to test efficiency (6586) fkiraly
[ENH] parametrize_with_checks utility for granular API compliance test setup in 2nd/3rd party libraries (6588) fkiraly
[ENH] differential testing to utils module (6620) fkiraly
[ENH] differential testing and minor improvements to forecasting.base tests (6619) fkiraly
[ENH] differential testing for performance_metrics module (6616) fkiraly
[ENH] fixes and improvements to pytest doctest integration (6621) fkiraly
[DOC] fix broken links on webpage docs (6339) duydl
[DOC] document more tags (6496) fkiraly
[DOC] fix minor typos in tags API reference (6631) fkiraly
[DOC] update dependencies reference (6655) emmanuel-ferdman
[DOC] fix minor typo in developer comment in BaseTransformer (6689) Spinachboul
[DOC] rst roadmap documentation page stale since 2021 - replace by correct links to recent roadmaps (6556) fkiraly
[DOC] clarify docs on ARIMA estimators, add author credits for upstream (6705) fkiraly
[DOC] credit doberbauer` for pykalman python 3.11 compatibility fix (6662) doberbauer
[MNT] [Dependabot](deps): Update holidays requirement from <0.51,>=0.29 to >=0.29,<0.52 (6634) dependabot[bot]
[MNT] [Dependabot](deps): Update holidays requirement from <0.52,>=0.29 to >=0.52,<0.53 (6702) dependabot[bot]
[MNT] [Dependabot](deps): Update dask requirement from <2024.6.1 to <2024.6.2 (6643) dependabot[bot]
[MNT] [Dependabot](deps): Update numba requirement from <0.60,>=0.53 to >=0.53,<0.61 (6590) dependabot[bot]
[MNT] [Dependabot](deps): Update dask requirement from <2024.6.2 to <2024.6.3 (6647) dependabot[bot]
[MNT] remove coverage reporting and pytest-cov from PR CI and setup.cfg (6363) fkiraly
[MNT] numpy 2 compatibility fixes - estimators (6626) fkiraly
[MNT] scipy 1.14.0 compatibility for deep_equals plugin for csr_matrix (6664) fkiraly
[MNT] deprecate unused _check_soft_dependencies argument suppress_import_stdout (6691) fkiraly
[BUG] fix AUCalibration probabilistic metric for multivariate case (6617) fkiraly
[BUG] fix bug 4076: PerformanceWarning in load_from_tsfile_to_dataframe (6632) ericjb
[BUG] patch over pandas 2.2.X issue in freq timestamp/period round trip conversion for period start timestamps such as "MonthBegin" (6574) fkiraly
[BUG] fix passing of y in ForecastingPipeline (6706) fkiraly
[BUG] Fix bug in fitted parameter override in pyts and tslearn adapters (6707) fkiraly
[BUG] Fix bug in fitted parameter override in pyts and tslearn adapters (6707) fkiraly
[BUG] in TimeSeriesForestRegressor, fix failure: self.criterion does not exist (6573) ksharma6
[BUG] partially revert pytest.skip change from #6233 due to side effects in downstream test suites (6508) fkiraly
[BUG] fix test failures introduced by differential testing refactor (6585) fkiraly
[BUG] fix HolidayFeatures crashes if dataframe doesn't contain specified date (6550) fnhirwa
[BUG] in Differencer, make explicit clone to avoid SettingWithCopyWarning (6567) benHeid
[BUG] minirocket: fix zero division errors #5174 (6612) benshaw2
[BUG] ensure correct setting of requires_X and requires_y tag for FeatureUnion (6695) fkiraly
[BUG] ensure correct setting of requires_X and requires_y tag for TransformerPipeline (6692) fkiraly
[BUG] partial fix for dropped column names in PaddingTransformer (6693) fkiraly
Abhay-Lejith, benHeid, benshaw2, doberbauer, emmanuel-ferdman, ericjb, felipeangelimvieira, fkiraly, fnhirwa, gareth-brown-86, geetu040, iaryangoyal, julian-fong, ksharma6, mk406, shlok191, Spinachboul, TheoWeih, XinyuWuu, yarnabrina, Z-Fran
Minimal maintenance patch improving onboard package structure without breaking changes.
Minimal maintenance update with actions consolidating onboard packages.
For last major feature update, see 0.29.1.
[MNT] reorganization of onboard libs - pykalman , vmdpy ( #6535 ) @fkiraly
[MNT] differential testing for split module ( #6532 ) @fkiraly
Minimal maintenance update with actions consolidating onboard packages.
For last major feature update, see 0.29.1.
[MNT] reorganization of onboard libs - pykalman, vmdpy (6535) fkiraly
[MNT] differential testing for split module (6532) fkiraly
Maintenance release and some breaking changes.
Maintenance release and some breaking changes.
For last larger feature update, see 0.29.1.
Please see our changelog for a description of all changes.
@Alex-JG3, @fkiraly, @gareth-brown-86, @geetu040, @yarnabrina
Full Changelog: https://github.com/sktime/sktime/compare/v0.29.1...v0.30.0
Major upgrade to the time series anomaly, changepoints, segmentation API ( @Alex-JG3 ). Users should review the section in the release notes.
Kindly also note the python 3.8 End-of-life warning below.
Also includes scheduled deprecations and change actions.
For last major feature update, see 0.29.1.
joblib is now an explicit core dependency, with bounds <1.5,>=1.2.0 . Previously, joblib was an indirect core dependency, via scikit-learn . Due to direct imports, this was changed to an explicit dependency.
scikit-learn (core dependency) bounds have been updated to >=0.24,<1.6.0
scikit-base (core dependency) bounds have been updated to >=0.6.1,<0.9.0
skpro (soft dependency) bounds have been updated to >=2,<2.4.0
kotsu is not longer a soft dependency required by the forecasting benchmarking framework. The kotsu package is no longer maintained, and its necessary imports have beend moved to sktime as private utilities until refactor. See #6514 .
pykalman (transformations soft dependency) has been forked into sktime , as sktime.libs.pykalman , as the original package is no longer maintained, see sktime issue 5414 or pykalman issue 109.
The package fork will be maintained in sktime .
Direct users of pykalman can replace imports from pykalman import x with equivalent imports from sktime.libs.pykalman import x .
Indirect users via the transformer KalmanFilterTransformerPK will not be impacted as APIs do not change, except that they no longer require the original pykalman package in their python environment.
The time series annotation, anomalies, changepoints, segmentation API has been fully reworked to be in line with scikit-base patterns, sktime tags, and to provide a more consistent and flexible interface.
the API provides predict methods for annotation labels, e.g., segments, outlier points, and a transform method for indicator series, for instance 1/0 indicator whether an anomaly is present at the time stamp.
the fmt argument used in some estimators is now deprecated, in favour of using predict or transform .
The type of annotation, e.g., change points or segmentation, is encoded by the new tag task used in time series annotators, with values anomaly_detection , segmentation , changepoint_detection .
Low-level methods allow polymorphic use of annotators, e.g., a changepoint detector to be used for segmentation, via predict_points or predict_segments . The predict method defaults to the type of annotation defined by task .
A full tutorial with examples will be created over the next release cycles, and further enhancements are planned.
sktime now requires Python version >=3.9 . No errors will be raised on Python 3.8, but test coverage and support for Python 3.8 has been dropped.
Kindly note for context: python 3.8 will reach end of life in October 2024, and multiple sktime core dependencies, including scikit-learn , have already dropped support for 3.8.
cINNForecaster has been renamed to CINNForecaster . The estimator is no longer available under its old name, after the deprecation period. Users should replace any imports of cINNForecaster with imports of CINNForecaster .
[ENH] Rework of base series annotator API ( #6265 ) @Alex-JG3
[ENH] upgrade is_module_changed test utility for paths ( #6518 ) @fkiraly
[DOC] updated all_estimators docstring for re.Pattern support ( #6478 ) @fkiraly
[MNT] Dependabot: Update skpro requirement from <2.3.0,>=2 to >=2,<2.4.0 ( #6443 ) @dependabot[bot]
[MNT] Dependabot: Update scikit-learn requirement from <1.5.0,>=0.24 to >=0.24,<1.6.0 ( #6462 ) @dependabot[bot]
[MNT] Dependabot: Update scikit-base requirement from <0.8.0,>=0.6.1 to >=0.6.1,<0.9.0 ( #6488 ) @dependabot[bot]
[MNT] drop test coverage on python 3.8 in CI ( #6329 ) @yarnabrina
[MNT] final change cycle (0.30.0) for renaming cINNForecaster to CINNForecaster ( #6367 ) @geetu040
[MNT] added joblib as core dependency ( #6384 ) @yarnabrina
[MNT] 0.30.0 deprecations and change actions ( #6468 ) @fkiraly
[MNT] modified CRLF line endings to LF line endings ( #6512 ) @yarnabrina
[MNT] Move dependency checkers to separate module in utils ( #6354 ) @fkiraly
[MNT] resolution to pykalman issue - sktime local pykalman fork ( #6188 ) @fkiraly
[MNT] add systematic differential test switch to low-level tests ( #6511 ) @fkiraly
[MNT] isolate utils module init and sktime init from external imports ( #6516 ) @fkiraly
[MNT] preparing refactor of benchmark framework: folding minimal kotsu library into sktime ( #6514 ) @fkiraly
[MNT] run tests in distances module only if it has changed ( #6517 ) @fkiraly
[MNT] refactor pykalman tests to pytest and conditional execution ( #6519 ) @fkiraly
[MNT] conditional execution of tests in datatypes module ( #6520 ) @fkiraly
@Alex-JG3 , @dependabot[bot] , @fkiraly , @geetu040 , @yarnabrina
Major upgrade to the time series anomaly, changepoints, segmentation API (Alex-JG3). Users should review the section in the release notes.
Kindly also note the python 3.8 End-of-life warning below.
Also includes scheduled deprecations and change actions.
For last major feature update, see 0.29.1.
joblib is now an explicit core dependency, with bounds <1.5,>=1.2.0. Previously, joblib was an indirect core dependency, via scikit-learn. Due to direct imports, this was changed to an explicit dependency.
scikit-learn (core dependency) bounds have been updated to >=0.24,<1.6.0
scikit-base (core dependency) bounds have been updated to >=0.6.1,<0.9.0
skpro (soft dependency) bounds have been updated to >=2,<2.4.0
kotsu is not longer a soft dependency required by the forecasting benchmarking framework. The kotsu package is no longer maintained, and its necessary imports have beend moved to sktime as private utilities until refactor. See 6514.
pykalman (transformations soft dependency) has been forked into sktime, as sktime.libs.pykalman, as the original package is no longer maintained, see sktime issue 5414 or pykalman issue 109.
The package fork will be maintained in sktime.
Direct users of pykalman can replace imports from pykalman import x with equivalent imports from sktime.libs.pykalman import x.
Indirect users via the transformer KalmanFilterTransformerPK will not be impacted as APIs do not change, except that they no longer require the original pykalman package in their python environment.
The time series annotation, anomalies, changepoints, segmentation API has been fully reworked to be in line with scikit-base patterns, sktime tags, and to provide a more consistent and flexible interface.
the API provides predict methods for annotation labels, e.g., segments, outlier points, and a transform method for indicator series, for instance 1/0 indicator whether an anomaly is present at the time stamp.
the fmt argument used in some estimators is now deprecated, in favour of using predict or transform.
The type of annotation, e.g., change points or segmentation, is encoded by the new tag task used in time series annotators, with values anomaly_detection, segmentation, changepoint_detection.
Low-level methods allow polymorphic use of annotators, e.g., a changepoint detector to be used for segmentation, via predict_points or predict_segments. The predict method defaults to the type of annotation defined by task.
A full tutorial with examples will be created over the next release cycles, and further enhancements are planned.
sktime now requires Python version >=3.9. No errors will be raised on Python 3.8, but test coverage and support for Python 3.8 has been dropped.
Kindly note for context: python 3.8 will reach end of life in October 2024, and multiple sktime core dependencies, including scikit-learn, have already dropped support for 3.8.
cINNForecaster has been renamed to CINNForecaster. The estimator is no longer available under its old name, after the deprecation period. Users should replace any imports of cINNForecaster with imports of CINNForecaster.
[ENH] Rework of base series annotator API (6265) Alex-JG3
[ENH] upgrade is_module_changed test utility for paths (6518) fkiraly
[DOC] updated all_estimators docstring for re.Pattern support (6478) fkiraly
[MNT] [Dependabot](deps): Update skpro requirement from <2.3.0,>=2 to >=2,<2.4.0 (6443) dependabot[bot]
[MNT] [Dependabot](deps): Update scikit-learn requirement from <1.5.0,>=0.24 to >=0.24,<1.6.0 (6462) dependabot[bot]
[MNT] [Dependabot](deps): Update scikit-base requirement from <0.8.0,>=0.6.1 to >=0.6.1,<0.9.0 (6488) dependabot[bot]
[MNT] drop test coverage on python 3.8 in CI (6329) yarnabrina
[MNT] final change cycle (0.30.0) for renaming cINNForecaster to CINNForecaster (6367) geetu040
[MNT] added joblib as core dependency (6384) yarnabrina
[MNT] 0.30.0 deprecations and change actions (6468) fkiraly
[MNT] modified CRLF line endings to LF line endings (6512) yarnabrina
[MNT] Move dependency checkers to separate module in utils (6354) fkiraly
[MNT] resolution to pykalman issue - sktime local pykalman fork (6188) fkiraly
[MNT] add systematic differential test switch to low-level tests (6511) fkiraly
[MNT] isolate utils module init and sktime init from external imports (6516) fkiraly
[MNT] preparing refactor of benchmark framework: folding minimal kotsu library into sktime (6514) fkiraly
[MNT] run tests in distances module only if it has changed (6517) fkiraly
[MNT] refactor pykalman tests to pytest and conditional execution (6519) fkiraly
[MNT] conditional execution of tests in datatypes module (6520) fkiraly
Alex-JG3, dependabot[bot], fkiraly, geetu040, yarnabrina
One column per quarter.