Version 1.6

Legend for changelogs

• Major Feature something big that you couldn’t do before.

• Feature something that you couldn’t do before.

• Efficiency an existing feature now may not require as much computation or memory.

• Enhancement a miscellaneous minor improvement.

• Fix something that previously didn’t work as documented – or according to reasonable expectations – should now work.

• API Change you will need to change your code to have the same effect in the future; or a feature will be removed in the future.

Version 1.6.0

In Development

Changes impacting many modules

• API Change utils.validation.validate_data is introduced and replaces previously private base.BaseEstimator._validate_data method. This is intended for third party estimator developers, who should use this function in most cases instead of utils.validation.check_array and utils.validation.check_X_y. #29696 by Adrin Jalali.

• Enhancement __sklearn_tags__ was introduced for setting tags in estimators. More details in Estimator Tags. #22606 by Thomas Fan and #29677 by Adrin Jalali.

Support for Array API

Additional estimators and functions have been updated to include support for all Array API compliant inputs.

See Array API support (experimental) for more details.

Functions:

• sklearn.metrics.cluster.entropy #29141 by Yaroslav Korobko;

• sklearn.metrics.d2_tweedie_score #29207 by Emily Chen;

• sklearn.metrics.max_error #29212 by Edoardo Abati;

• sklearn.metrics.mean_absolute_error #27736 by Edoardo Abati and #29143 by Tialo and Loïc Estève;

• sklearn.metrics.mean_absolute_percentage_error #29300 by Emily Chen;

• sklearn.metrics.mean_gamma_deviance #29239 by Emily Chen;

• sklearn.metrics.mean_poisson_deviance #29227 by Emily Chen;

• sklearn.metrics.mean_squared_error #29142 by Yaroslav Korobko;

• sklearn.metrics.mean_squared_log_error #29709 by Virgil Chan;

• sklearn.metrics.mean_tweedie_deviance #28106 by Thomas Li;

• sklearn.metrics.root_mean_squared_error #29709 by Virgil Chan;

• sklearn.metrics.root_mean_squared_log_error #29709 by Virgil Chan;

• sklearn.metrics.pairwise.additive_chi2_kernel #29144 by Yaroslav Korobko;

• sklearn.metrics.pairwise.chi2_kernel #29267 by Yaroslav Korobko;

• sklearn.metrics.pairwise.cosine_distances #29265 by Emily Chen;

• sklearn.metrics.pairwise.cosine_similarity #29014 by Edoardo Abati;

• sklearn.metrics.pairwise.euclidean_distances #29433 by Omar Salman;

• sklearn.metrics.pairwise.linear_kernel #29475 by Omar Salman;

• sklearn.metrics.pairwise.paired_cosine_distances #29112 by Edoardo Abati.

• sklearn.metrics.pairwise.paired_euclidean_distances #29389 by Emily Chen;

• sklearn.metrics.pairwise.polynomial_kernel #29475 by Omar Salman;

• sklearn.metrics.pairwise.rbf_kernel #29433 by Omar Salman;

• sklearn.metrics.pairwise.sigmoid_kernel #29475 by Omar Salman.

Classes:

• preprocessing.LabelEncoder now supports Array API compatible inputs. #27381 by Omar Salman.

• model_selection.GridSearchCV, model_selection.RandomizedSearchCV, model_selection.HalvingGridSearchCV and model_selection.HalvingRandomSearchCV now support Array API compatible inputs when their base estimators do. #27096 by Tim Head and Olivier Grisel.

• preprocessing.MinMaxScaler with clip=True #29751 by Shreekant Nandiyawar

Other

• Support for the soon to be deprecated cupy.array_api module has been removed in favor of directly supporting the top level cupy module, possibly via the array_api_compat.cupy compatibility wrapper. #29639 by Olivier Grisel.

Metadata Routing

The following models now support metadata routing in one or more of their methods. Refer to the Metadata Routing User Guide for more details.

• Feature model_selection.learning_curve now supports metadata routing for the fit method of its estimator and for its underlying CV splitter and scorer. #28975 by Stefanie Senger.

• Feature ensemble.StackingClassifier and ensemble.StackingRegressor now support metadata routing and pass **fit_params to the underlying estimators via their fit methods. #28701 by Stefanie Senger.

• Feature compose.TransformedTargetRegressor now supports metadata routing in its fit and predict methods and routes the corresponding params to the underlying regressor. #29136 by Omar Salman.

• Feature feature_selection.SequentialFeatureSelector now supports metadata routing in its fit method and passes the corresponding params to the model_selection.cross_val_score function. #29260 by Omar Salman.

• Feature model_selection.validation_curve now supports metadata routing for the fit method of its estimator and for its underlying CV splitter and scorer. #29329 by Stefanie Senger.

• Feature semi_supervised.SelfTrainingClassifier now supports metadata routing. The fit method now accepts **fit_params which are passed to the underlying estimators via their fit methods. In addition, the predict, predict_proba, predict_log_proba, score and decision_function methods also accept **params which are passed to the underlying estimators via their respective methods. #28494 by Adam Li.

• Feature model_selection.permutation_test_score now supports metadata routing for the fit method of its estimator and for its underlying CV splitter and scorer. #29266 by Adam Li.

• Feature feature_selection.RFE and feature_selection.RFECV now support metadata routing. #29312 by Omar Salman.

• Fix Metadata is routed correctly to grouped CV splitters via linear_model.RidgeCV and linear_model.RidgeClassifierCV and UnsetMetadataPassedError is fixed for linear_model.RidgeClassifierCV with default scoring. #29634 by Stefanie Senger.

Dropping support for building with setuptools

From scikit-learn 1.6 onwards, support for building with setuptools has been removed. Meson is the only supported way to build scikit-learn, see Building from source for more details.

#29400 by Loïc Estève

Dropping official support for PyPy

Due to limited maintainer resources and small number of users, official PyPy support has been dropped. Some parts of scikit-learn may still work but PyPy is not tested anymore in the scikit-learn Continuous Integration. #29128 by Loïc Estève.

Changelog

sklearn.base

• Enhancement Added a function base.is_clusterer which determines whether a given estimator is of category clusterer. #28936 by Christian Veenhuis.

sklearn.cluster

• API Change The copy parameter of cluster.Birch was deprecated in 1.6 and will be removed in 1.8. It has no effect as the estimator does not perform in-place operations on the input data. #29124 by Yao Xiao.

sklearn.compose

• Enhancement sklearn.compose.ColumnTransformer verbose_feature_names_out now accepts string format or callable to generate feature names. #28934 by Marc Bresson.

sklearn.covariance

• Efficiency covariance.MinCovDet fitting is now slightly faster. #29835 by Antony Lee.

sklearn.cross_decomposition

• Fix cross_decomposition.PLSRegression properly raises an error when n_components is larger than n_samples. #29710 by Thomas Fan.

sklearn.datasets

• Feature datasets.fetch_file allows downloading arbitrary data-file from the web. It handles local caching, integrity checks with SHA256 digests and automatic retries in case of HTTP errors. #29354 by Olivier Grisel.

sklearn.discriminant_analysis

• Fix discriminant_analysis.QuadraticDiscriminantAnalysis will now cause LinAlgWarning in case of collinear variables. These errors can be silenced using the reg_param attribute. #19731 by Alihan Zihna.

sklearn.ensemble

• Efficiency Small runtime improvement of fitting ensemble.HistGradientBoostingClassifier and ensemble.HistGradientBoostingRegressor by parallelizing the initial search for bin thresholds. #28064 by Christian Lorentzen.

• Enhancement The verbosity of ensemble.HistGradientBoostingClassifier and ensemble.HistGradientBoostingRegressor got a more granular control. Now, verbose = 1 prints only summary messages, verbose >= 2 prints the full information as before. #28179 by Christian Lorentzen.

• Efficiency ensemble.IsolationForest now runs parallel jobs during predict offering a speedup of up to 2-4x on sample sizes larger than 2000 using joblib. #28622 by Adam Li and Sérgio Pereira.

• Feature ensemble.ExtraTreesClassifier and ensemble.ExtraTreesRegressor now support missing-values in the data matrix X. Missing-values are handled by randomly moving all of the samples to the left, or right child node as the tree is traversed. #28268 by Adam Li.

• API Change The parameter algorithm of ensemble.AdaBoostClassifier is deprecated and will be removed in 1.8. #29997 by Jérémie du Boisberranger.

sklearn.impute

• Fix impute.KNNImputer excludes samples with nan distances when computing the mean value for uniform weights. #29135 by Xuefeng Xu.

sklearn.linear_model

• Fix linear_model.LogisticRegressionCV corrects sample weight handling for the calculation of test scores. #29419 by Shruti Nath.

• Fix linear_model.RidgeCV now properly supports custom multioutput scorers by letting the scorer manage the multioutput averaging. Previously, the predictions and true targets were both squeezed to a 1D array before computing the error. #29884 by Guillaume Lemaitre.

• Fix linear_model.RidgeCV now properly uses predictions on the same scale as the target seen during fit. These predictions are stored in cv_results_ when scoring != None. Previously, the predictions were rescaled by the square root of the sample weights and offset by the mean of the target, leading to an incorrect estimate of the score. #29842 by Guillaume Lemaitre, Jérôme Dockes and Hanmin Qin.

• API Change Deprecates copy_X in linear_model.TheilSenRegressor as the parameter has no effect. copy_X will be removed in 1.8. #29105 by Adam Li.

• Fix linear_model.LassoCV and linear_model.ElasticNetCV now take sample weights into accounts to define the search grid for the internally tuned alpha hyper-parameter. #29442 by John Hopfensperger <s-banach> and :user:`Shruti Nath.

sklearn.manifold

• Efficiency manifold.locally_linear_embedding and manifold.LocallyLinearEmbedding now allocate more efficiently the memory of sparse matrices in the Hessian, Modified and LTSA methods. #28096 by Giorgio Angelotti.

sklearn.metrics

• Enhancement sklearn.metrics.check_scoring now accepts raise_exc to specify whether to raise an exception if a subset of the scorers in multimetric scoring fails or to return an error code. #28992 by Stefanie Senger.

• Enhancement Adds zero_division to cohen_kappa_score. When there is a division by zero, the metric is undefined and this value is returned. #29210 by Marc Torrellas Socastro and Stefanie Senger.

• Efficiency sklearn.metrics.classification_report is now faster by caching classification labels. #29738 by Adrin Jalali.

• API Change scoring=”neg_max_error” should be used instead of scoring=”max_error” which is now deprecated. #29462 by Farid “Freddie” Taba.

• API Change the assert_all_finite parameter of functions metrics.pairwise.check_pairwise_arrays and metrics.pairwise_distances is renamed into ensure_all_finite. force_all_finite will be removed in 1.8. #29404 by Jérémie du Boisberranger.

• Fix the functions metrics.mean_squared_log_error and metrics.root_mean_squared_log_error now check whether the inputs are within the correct domain for the function \(y=\log(1+x)\), rather than \(y=\log(x)\). #29709 by Virgil Chan.

• Fix the functions metrics.mean_absolute_error, metrics.mean_absolute_percentage_error, metrics.mean_squared_error and metrics.root_mean_squared_error now explicitly check whether a scalar will be returned when multioutput=uniform_average. #29709 by Virgil Chan.

sklearn.model_selection

• Enhancement Add the parameter prefit to model_selection.FixedThresholdClassifier allowing the use of a pre-fitted estimator without re-fitting it. #29067 by Guillaume Lemaitre.

• Fix Improve error message when model_selection.RepeatedStratifiedKFold.split is called without a y argument #29402 by Anurag Varma.

sklearn.neighbors

• Fix neighbors.LocalOutlierFactor raises a warning in the fit method when duplicate values in the training data lead to inaccurate outlier detection. #28773 by Henrique Caroço.

sklearn.preprocessing

• Enhancement The HTML representation of preprocessing.FunctionTransformer will show the function name in the label. #29158 by Yao Xiao.

• Fix preprocessing.PowerTransformer now uses scipy.special.inv_boxcox to output nan if the input of BoxCox’s inverse is invalid. #27875 by Xuefeng Xu.

sklearn.semi_supervised

• API Change semi_supervised.SelfTrainingClassifier deprecated the base_estimator parameter in favor of estimator. #28494 by Adam Li.

sklearn.tree

• Feature tree.ExtraTreeClassifier and tree.ExtraTreeRegressor now support missing-values in the data matrix X. Missing-values are handled by randomly moving all of the samples to the left, or right child node as the tree is traversed. #27966 by Adam Li.

sklearn.utils

• Enhancement utils.validation.check_array now accepts ensure_non_negative to check for negative values in the passed array, until now only available through calling utils.validation.check_non_negative. #29540 by Tamara Atanasoska.

• Fix utils.estimator_checks.parametrize_with_checks and utils.estimator_checks.check_estimator now support estimators that have set_output called on them. #29869 by Adrin Jalali.

• Enhancement utils.validation.check_is_fitted now passes on stateless estimators. An estimator can indicate it’s stateless by setting the requires_fit tag. See Estimator Tags for more information. #29880 by Adrin Jalali.

• API Change the assert_all_finite parameter of functions utils.check_array, utils.check_X_y, utils.as_float_array is renamed into ensure_all_finite. force_all_finite will be removed in 1.8. #29404 by Jérémie du Boisberranger.

Code and documentation contributors

Thanks to everyone who has contributed to the maintenance and improvement of the project since version 1.5, including:

TODO: update at the time of the release.