.. include:: _contributors.rst .. currentmodule:: sklearn .. _changes_1_2: Version 1.2.0 ============= **December 2022** For a short description of the main highlights of the release, please refer to :ref:`sphx_glr_auto_examples_release_highlights_plot_release_highlights_1_2_0.py`. .. include:: changelog_legend.inc Changed models -------------- The following estimators and functions, when fit with the same data and parameters, may produce different models from the previous version. This often occurs due to changes in the modelling logic (bug fixes or enhancements), or in random sampling procedures. - |Enhancement| The default `eigen_tol` for :class:`cluster.SpectralClustering`, :class:`manifold.SpectralEmbedding`, :func:`cluster.spectral_clustering`, and :func:`manifold.spectral_embedding` is now `None` when using the `'amg'` or `'lobpcg'` solvers. This change improves numerical stability of the solver, but may result in a different model. - |Enhancement| :class:`linear_model.GammaRegressor`, :class:`linear_model.PoissonRegressor` and :class:`linear_model.TweedieRegressor` can reach higher precision with the lbfgs solver, in particular when `tol` is set to a tiny value. Moreover, `verbose` is now properly propagated to L-BFGS-B. :pr:`23619` by :user:`Christian Lorentzen `. - |Enhancement| The default value for `eps` :func:`metrics.logloss` has changed from `1e-15` to `"auto"`. `"auto"` sets `eps` to `np.finfo(y_pred.dtype).eps`. :pr:`24354` by :user:`Safiuddin Khaja ` and :user:`gsiisg `. - |Fix| Make sign of `components_` deterministic in :class:`decomposition.SparsePCA`. :pr:`23935` by :user:`Guillaume Lemaitre `. - |Fix| The `components_` signs in :class:`decomposition.FastICA` might differ. It is now consistent and deterministic with all SVD solvers. :pr:`22527` by :user:`Meekail Zain ` and `Thomas Fan`_. - |Fix| The condition for early stopping has now been changed in :func:`linear_model._sgd_fast._plain_sgd` which is used by :class:`linear_model.SGDRegressor` and :class:`linear_model.SGDClassifier`. The old condition did not disambiguate between training and validation set and had an effect of overscaling the error tolerance. This has been fixed in :pr:`23798` by :user:`Harsh Agrawal `. - |Fix| For :class:`model_selection.GridSearchCV` and :class:`model_selection.RandomizedSearchCV` ranks corresponding to nan scores will all be set to the maximum possible rank. :pr:`24543` by :user:`Guillaume Lemaitre `. - |API| The default value of `tol` was changed from `1e-3` to `1e-4` for :func:`linear_model.ridge_regression`, :class:`linear_model.Ridge` and :class:`linear_model.`RidgeClassifier`. :pr:`24465` by :user:`Christian Lorentzen `. Changes impacting all modules ----------------------------- - |MajorFeature| The `set_output` API has been adopted by all transformers. Meta-estimators that contain transformers such as :class:`pipeline.Pipeline` or :class:`compose.ColumnTransformer` also define a `set_output`. For details, see `SLEP018 `__. :pr:`23734` and :pr:`24699` by `Thomas Fan`_. - |Efficiency| Low-level routines for reductions on pairwise distances for dense float32 datasets have been refactored. The following functions and estimators now benefit from improved performances in terms of hardware scalability and speed-ups: - :func:`sklearn.metrics.pairwise_distances_argmin` - :func:`sklearn.metrics.pairwise_distances_argmin_min` - :class:`sklearn.cluster.AffinityPropagation` - :class:`sklearn.cluster.Birch` - :class:`sklearn.cluster.MeanShift` - :class:`sklearn.cluster.OPTICS` - :class:`sklearn.cluster.SpectralClustering` - :func:`sklearn.feature_selection.mutual_info_regression` - :class:`sklearn.neighbors.KNeighborsClassifier` - :class:`sklearn.neighbors.KNeighborsRegressor` - :class:`sklearn.neighbors.RadiusNeighborsClassifier` - :class:`sklearn.neighbors.RadiusNeighborsRegressor` - :class:`sklearn.neighbors.LocalOutlierFactor` - :class:`sklearn.neighbors.NearestNeighbors` - :class:`sklearn.manifold.Isomap` - :class:`sklearn.manifold.LocallyLinearEmbedding` - :class:`sklearn.manifold.TSNE` - :func:`sklearn.manifold.trustworthiness` - :class:`sklearn.semi_supervised.LabelPropagation` - :class:`sklearn.semi_supervised.LabelSpreading` For instance :class:`sklearn.neighbors.NearestNeighbors.kneighbors` and :class:`sklearn.neighbors.NearestNeighbors.radius_neighbors` can respectively be up to ×20 and ×5 faster than previously on a laptop. Moreover, implementations of those two algorithms are now suitable for machine with many cores, making them usable for datasets consisting of millions of samples. :pr:`23865` by :user:`Julien Jerphanion `. - |Enhancement| Finiteness checks (detection of NaN and infinite values) in all estimators are now significantly more efficient for float32 data by leveraging NumPy's SIMD optimized primitives. :pr:`23446` by :user:`Meekail Zain ` - |Enhancement| Finiteness checks (detection of NaN and infinite values) in all estimators are now faster by utilizing a more efficient stop-on-first second-pass algorithm. :pr:`23197` by :user:`Meekail Zain ` - |Enhancement| Support for combinations of dense and sparse datasets pairs for all distance metrics and for float32 and float64 datasets has been added or has seen its performance improved for the following estimators: - :func:`sklearn.metrics.pairwise_distances_argmin` - :func:`sklearn.metrics.pairwise_distances_argmin_min` - :class:`sklearn.cluster.AffinityPropagation` - :class:`sklearn.cluster.Birch` - :class:`sklearn.cluster.SpectralClustering` - :class:`sklearn.neighbors.KNeighborsClassifier` - :class:`sklearn.neighbors.KNeighborsRegressor` - :class:`sklearn.neighbors.RadiusNeighborsClassifier` - :class:`sklearn.neighbors.RadiusNeighborsRegressor` - :class:`sklearn.neighbors.LocalOutlierFactor` - :class:`sklearn.neighbors.NearestNeighbors` - :class:`sklearn.manifold.Isomap` - :class:`sklearn.manifold.TSNE` - :func:`sklearn.manifold.trustworthiness` :pr:`23604` and :pr:`23585` by :user:`Julien Jerphanion `, :user:`Olivier Grisel `, and `Thomas Fan`_, :pr:`24556` by :user:`Vincent Maladière `. - |Fix| Systematically check the sha256 digest of dataset tarballs used in code examples in the documentation. :pr:`24617` by :user:`Olivier Grisel ` and `Thomas Fan`_. Thanks to `Sim4n6 `_ for the report. Changelog --------- .. Entries should be grouped by module (in alphabetic order) and prefixed with one of the labels: |MajorFeature|, |Feature|, |Efficiency|, |Enhancement|, |Fix| or |API| (see whats_new.rst for descriptions). Entries should be ordered by those labels (e.g. |Fix| after |Efficiency|). Changes not specific to a module should be listed under *Multiple Modules* or *Miscellaneous*. Entries should end with: :pr:`123456` by :user:`Joe Bloggs `. where 123456 is the *pull request* number, not the issue number. :mod:`sklearn.base` ................... - |Enhancement| Introduces :class:`base.ClassNamePrefixFeaturesOutMixin` and :class:`base.ClassNamePrefixFeaturesOutMixin` mixins that defines :term:`get_feature_names_out` for common transformer uses cases. :pr:`24688` by `Thomas Fan`_. :mod:`sklearn.calibration` .......................... - |API| Rename `base_estimator` to `estimator` in :class:`calibration.CalibratedClassifierCV` to improve readability and consistency. The parameter `base_estimator` is deprecated and will be removed in 1.4. :pr:`22054` by :user:`Kevin Roice `. :mod:`sklearn.cluster` ...................... - |Efficiency| :class:`cluster.KMeans` with `algorithm="lloyd"` is now faster and uses less memory. :pr:`24264` by :user:`Vincent Maladiere `. - |Enhancement| The `predict` and `fit_predict` methods of :class:`cluster.OPTICS` now accept sparse data type for input data. :pr:`14736` by :user:`Hunt Zhan `, :pr:`20802` by :user:`Brandon Pokorny `, and :pr:`22965` by :user:`Meekail Zain `. - |Enhancement| :class:`cluster.Birch` now preserves dtype for `numpy.float32` inputs. :pr:`22968` by `Meekail Zain `. - |Enhancement| :class:`cluster.KMeans` and :class:`cluster.MiniBatchKMeans` now accept a new `'auto'` option for `n_init` which changes the number of random initializations to one when using `init='k-means++'` for efficiency. This begins deprecation for the default values of `n_init` in the two classes and both will have their defaults changed to `n_init='auto'` in 1.4. :pr:`23038` by :user:`Meekail Zain `. - |Enhancement| :class:`cluster.SpectralClustering` and :func:`cluster.spectral_clustering` now propogates the `eigen_tol` parameter to all choices of `eigen_solver`. Includes a new option `eigen_tol="auto"` and begins deprecation to change the default from `eigen_tol=0` to `eigen_tol="auto"` in version 1.3. :pr:`23210` by :user:`Meekail Zain `. - |Fix| :class:`cluster.KMeans` now supports readonly attributes when predicting. :pr:`24258` by `Thomas Fan`_ - |API| The `affinity` attribute is now deprecated for :class:`cluster.AgglomerativeClustering` and will be renamed to `metric` in v1.4. :pr:`23470` by :user:`Meekail Zain `. :mod:`sklearn.datasets` ....................... - |Enhancement| Introduce the new parameter `parser` in :func:`datasets.fetch_openml`. `parser="pandas"` allows to use the very CPU and memory efficient `pandas.read_csv` parser to load dense ARFF formatted dataset files. It is possible to pass `parser="liac-arff"` to use the old LIAC parser. When `parser="auto"`, dense datasets are loaded with "pandas" and sparse datasets are loaded with "liac-arff". Currently, `parser="liac-arff"` by default and will change to `parser="auto"` in version 1.4 :pr:`21938` by :user:`Guillaume Lemaitre `. - |Enhancement| :func:`datasets.dump_svmlight_file` is now accelerated with a Cython implementation, providing 2-4x speedups. :pr:`23127` by :user:`Meekail Zain ` - |Enhancement| Path-like objects, such as those created with pathlib are now allowed as paths in :func:`datasets.load_svmlight_file` and :func:`datasets.load_svmlight_files`. :pr:`19075` by :user:`Carlos Ramos Carreño `. - |Fix| Make sure that :func:`datasets.fetch_lfw_people` and :func:`datasets.fetch_lfw_pairs` internally crops images based on the `slice_` parameter. :pr:`24951` by :user:`Guillaume Lemaitre `. :mod:`sklearn.decomposition` ............................ - |Efficiency| :func:`decomposition.FastICA.fit` has been optimised w.r.t its memory footprint and runtime. :pr:`22268` by :user:`MohamedBsh `. - |Enhancement| :class:`decomposition.SparsePCA` and :class:`decomposition.MiniBatchSparsePCA` now implements an `inverse_transform` function. :pr:`23905` by :user:`Guillaume Lemaitre `. - |Enhancement| :class:`decomposition.FastICA` now allows the user to select how whitening is performed through the new `whiten_solver` parameter, which supports `svd` and `eigh`. `whiten_solver` defaults to `svd` although `eigh` may be faster and more memory efficient in cases where `num_features > num_samples`. :pr:`11860` by :user:`Pierre Ablin `, :pr:`22527` by :user:`Meekail Zain ` and `Thomas Fan`_. - |Enhancement| :class:`decomposition.LatentDirichletAllocation` now preserves dtype for `numpy.float32` input. :pr:`24528` by :user:`Takeshi Oura ` and :user:`Jérémie du Boisberranger `. - |Fix| Make sign of `components_` deterministic in :class:`decomposition.SparsePCA`. :pr:`23935` by :user:`Guillaume Lemaitre `. - |API| The `n_iter` parameter of :class:`decomposition.MiniBatchSparsePCA` is deprecated and replaced by the parameters `max_iter`, `tol`, and `max_no_improvement` to be consistent with :class:`decomposition.MiniBatchDictionaryLearning`. `n_iter` will be removed in version 1.3. :pr:`23726` by :user:`Guillaume Lemaitre `. - |API| The `n_features_` attribute of :class:`decomposition.PCA` is deprecated in favor of `n_features_in_` and will be removed in 1.4. :pr:`24421` by :user:`Kshitij Mathur `. :mod:`sklearn.discriminant_analysis` .................................... - |MajorFeature| :class:`discriminant_analysis.LinearDiscriminantAnalysis` now supports the `Array API `_ for `solver="svd"`. Array API support is considered experimental and might evolve without being subjected to our usual rolling deprecation cycle policy. See :ref:`array_api` for more details. :pr:`22554` by `Thomas Fan`_. - |Fix| Validate parameters only in `fit` and not in `__init__` for :class:`discriminant_analysis.QuadraticDiscriminantAnalysis`. :pr:`24218` by :user:`Stefanie Molin `. :mod:`sklearn.ensemble` ....................... - |MajorFeature| :class:`ensemble.HistGradientBoostingClassifier` and :class:`ensemble.HistGradientBoostingRegressor` now support interaction constraints via the argument `interaction_cst` of their constructors. :pr:`21020` by :user:`Christian Lorentzen `. Using interaction constraints also makes fitting faster. :pr:`24856` by :user:`Christian Lorentzen `. - |Feature| Adds `class_weight` to :class:`ensemble.HistGradientBoostingClassifier`. :pr:`22014` by `Thomas Fan`_. - |Efficiency| Improve runtime performance of :class:`ensemble.IsolationForest` by avoiding data copies. :pr:`23252` by :user:`Zhehao Liu `. - |Enhancement| :class:`ensemble.StackingClassifier` now accepts any kind of base estimator. :pr:`24538` by :user:`Guillem G Subies `. - |Enhancement| Make it possible to pass the `categorical_features` parameter of :class:`ensemble.HistGradientBoostingClassifier` and :class:`ensemble.HistGradientBoostingRegressor` as feature names. :pr:`24889` by :user:`Olivier Grisel `. - |Enhancement| :class:`ensemble.StackingClassifier` now supports multilabel-indicator target :pr:`24146` by :user:`Nicolas Peretti `, :user:`Nestor Navarro `, :user:`Nati Tomattis `, and :user:`Vincent Maladiere `. - |Enhancement| :class:`ensemble.HistGradientBoostingClassifier` and :class:`ensemble.HistGradientBoostingClassifier` now accept their `monotonic_cst` parameter to be passed as a dictionary in addition to the previously supported array-like format. Such dictionary have feature names as keys and one of `-1`, `0`, `1` as value to specify monotonicity constraints for each feature. :pr:`24855` by :user:`Olivier Grisel `. - |Enhancement| Interaction constraints for :class:`ensemble.HistGradientBoostingClassifier` and :class:`ensemble.HistGradientBoostingRegressor` can now be specified as strings for two common cases: "no_interactions" and "pairwise" interactions. :pr:`24849` by :user:`Tim Head `. - |Fix| Fixed the issue where :class:`ensemble.AdaBoostClassifier` outputs NaN in feature importance when fitted with very small sample weight. :pr:`20415` by :user:`Zhehao Liu `. - |Fix| :class:`ensemble.HistGradientBoostingClassifier` and :class:`ensemble.HistGradientBoostingRegressor` no longer error when predicting on categories encoded as negative values and instead consider them a member of the "missing category". :pr:`24283` by `Thomas Fan`_. - |Fix| :class:`ensemble.HistGradientBoostingClassifier` and :class:`ensemble.HistGradientBoostingRegressor`, with `verbose>=1`, print detailed timing information on computing histograms and finding best splits. The time spent in the root node was previously missing and is now included in the printed information. :pr:`24894` by :user:`Christian Lorentzen `. - |API| Rename the constructor parameter `base_estimator` to `estimator` in the following classes: :class:`ensemble.BaggingClassifier`, :class:`ensemble.BaggingRegressor`, :class:`ensemble.AdaBoostClassifier`, :class:`ensemble.AdaBoostRegressor`. `base_estimator` is deprecated in 1.2 and will be removed in 1.4. :pr:`23819` by :user:`Adrian Trujillo ` and :user:`Edoardo Abati `. - |API| Rename the fitted attribute `base_estimator_` to `estimator_` in the following classes: :class:`ensemble.BaggingClassifier`, :class:`ensemble.BaggingRegressor`, :class:`ensemble.AdaBoostClassifier`, :class:`ensemble.AdaBoostRegressor`, :class:`ensemble.RandomForestClassifier`, :class:`ensemble.RandomForestRegressor`, :class:`ensemble.ExtraTreesClassifier`, :class:`ensemble.ExtraTreesRegressor`, :class:`ensemble.RandomTreesEmbedding`, :class:`ensemble.IsolationForest`. `base_estimator_` is deprecated in 1.2 and will be removed in 1.4. :pr:`23819` by :user:`Adrian Trujillo ` and :user:`Edoardo Abati `. :mod:`sklearn.feature_selection` ................................ - |Fix| Fix a bug in :func:`feature_selection.mutual_info_regression` and :func:`feature_selection.mutual_info_classif`, where the continuous features in `X` should be scaled to a unit variance independently if the target `y` is continuous or discrete. :pr:`24747` by :user:`Guillaume Lemaitre ` :mod:`sklearn.gaussian_process` ............................... - |Fix| Fix :class:`gaussian_process.kernels.Matern` gradient computation with `nu=0.5` for PyPy (and possibly other non CPython interpreters). :pr:`24245` by :user:`Loïc Estève `. - |Fix| The `fit` method of :class:`gaussian_process.GaussianProcessRegressor` will not modify the input X in case a custom kernel is used, with a `diag` method that returns part of the input X. :pr:`24405` by :user:`Omar Salman `. :mod:`sklearn.impute` ..................... - |Enhancement| Added `keep_empty_features` parameter to :class:`impute.SimpleImputer`, :class:`impute.KNNImputer` and :class:`impute.IterativeImputer`, preventing removal of features containing only missing values when transforming. :pr:`16695` by :user:`Vitor Santa Rosa `. :mod:`sklearn.inspection` ......................... - |MajorFeature| Extended :func:`inspection.partial_dependence` and :class:`inspection.PartialDependenceDisplay` to handle categorical features. :pr:`18298` by :user:`Madhura Jayaratne ` and :user:`Guillaume Lemaitre `. - |Fix| :class:`inspection.DecisionBoundaryDisplay` now raises error if input data is not 2-dimensional. :pr:`25077` by :user:`Arturo Amor `. :mod:`sklearn.kernel_approximation` ................................... - |Enhancement| :class:`kernel_approximation.RBFSampler` now preserves dtype for `numpy.float32` inputs. :pr:`24317` by `Tim Head `. - |Enhancement| :class:`kernel_approximation.SkewedChi2Sampler` now preserves dtype for `numpy.float32` inputs. :pr:`24350` by :user:`Rahil Parikh `. - |Enhancement| :class:`kernel_approximation.RBFSampler` now accepts `'scale'` option for parameter `gamma`. :pr:`24755` by :user:`Gleb Levitski `. :mod:`sklearn.linear_model` ........................... - |Enhancement| :class:`linear_model.LogisticRegression`, :class:`linear_model.LogisticRegressionCV`, :class:`linear_model.GammaRegressor`, :class:`linear_model.PoissonRegressor` and :class:`linear_model.TweedieRegressor` got a new solver `solver="newton-cholesky"`. This is a 2nd order (Newton) optimisation routine that uses a Cholesky decomposition of the hessian matrix. When `n_samples >> n_features`, the `"newton-cholesky"` solver has been observed to converge both faster and to a higher precision solution than the `"lbfgs"` solver on problems with one-hot encoded categorical variables with some rare categorical levels. :pr:`24637` and :pr:`24767` by :user:`Christian Lorentzen `. - |Enhancement| :class:`linear_model.GammaRegressor`, :class:`linear_model.PoissonRegressor` and :class:`linear_model.TweedieRegressor` can reach higher precision with the lbfgs solver, in particular when `tol` is set to a tiny value. Moreover, `verbose` is now properly propagated to L-BFGS-B. :pr:`23619` by :user:`Christian Lorentzen `. - |Fix| :class:`linear_model.SGDClassifier` and :class:`linear_model.SGDRegressor` will raise an error when all the validation samples have zero sample weight. :pr:`23275` by `Zhehao Liu `. - |Fix| :class:`linear_model.SGDOneClassSVM` no longer performs parameter validation in the constructor. All validation is now handled in `fit()` and `partial_fit()`. :pr:`24433` by :user:`Yogendrasingh `, :user:`Arisa Y. ` and :user:`Tim Head `. - |Fix| Fix average loss calculation when early stopping is enabled in :class:`linear_model.SGDRegressor` and :class:`linear_model.SGDClassifier`. Also updated the condition for early stopping accordingly. :pr:`23798` by :user:`Harsh Agrawal `. - |API| The default value for the `solver` parameter in :class:`linear_model.QuantileRegressor` will change from `"interior-point"` to `"highs"` in version 1.4. :pr:`23637` by :user:`Guillaume Lemaitre `. - |API| String option `"none"` is deprecated for `penalty` argument in :class:`linear_model.LogisticRegression`, and will be removed in version 1.4. Use `None` instead. :pr:`23877` by :user:`Zhehao Liu `. - |API| The default value of `tol` was changed from `1e-3` to `1e-4` for :func:`linear_model.ridge_regression`, :class:`linear_model.Ridge` and :class:`linear_model.RidgeClassifier`. :pr:`24465` by :user:`Christian Lorentzen `. :mod:`sklearn.manifold` ....................... - |Feature| Adds option to use the normalized stress in :class:`manifold.MDS`. This is enabled by setting the new `normalize` parameter to `True`. :pr:`10168` by :user:`Łukasz Borchmann `, :pr:`12285` by :user:`Matthias Miltenberger `, :pr:`13042` by :user:`Matthieu Parizy `, :pr:`18094` by :user:`Roth E Conrad ` and :pr:`22562` by :user:`Meekail Zain `. - |Enhancement| Adds `eigen_tol` parameter to :class:`manifold.SpectralEmbedding`. Both :func:`manifold.spectral_embedding` and :class:`manifold.SpectralEmbedding` now propogate `eigen_tol` to all choices of `eigen_solver`. Includes a new option `eigen_tol="auto"` and begins deprecation to change the default from `eigen_tol=0` to `eigen_tol="auto"` in version 1.3. :pr:`23210` by :user:`Meekail Zain `. - |Enhancement| :class:`manifold.Isomap` now preserves dtype for `np.float32` inputs. :pr:`24714` by :user:`Rahil Parikh `. - |API| Added an `"auto"` option to the `normalized_stress` argument in :class:`manifold.MDS` and :func:`manifold.smacof`. Note that `normalized_stress` is only valid for non-metric MDS, therefore the `"auto"` option enables `normalized_stress` when `metric=False` and disables it when `metric=True`. `"auto"` will become the default value foor `normalized_stress` in version 1.4. :pr:`23834` by :user:`Meekail Zain ` :mod:`sklearn.metrics` ...................... - |Feature| :func:`metrics.ConfusionMatrixDisplay.from_estimator`, :func:`metrics.ConfusionMatrixDisplay.from_predictions`, and :meth:`metrics.ConfusionMatrixDisplay.plot` accepts a `text_kw` parameter which is passed to matplotlib's `text` function. :pr:`24051` by `Thomas Fan`_. - |Feature| :func:`metrics.class_likelihood_ratios` is added to compute the positive and negative likelihood ratios derived from the confusion matrix of a binary classification problem. :pr:`22518` by :user:`Arturo Amor `. - |Feature| Add :class:`metrics.PredictionErrorDisplay` to plot residuals vs predicted and actual vs predicted to qualitatively assess the behavior of a regressor. The display can be created with the class methods :func:`metrics.PredictionErrorDisplay.from_estimator` and :func:`metrics.PredictionErrorDisplay.from_predictions`. :pr:`18020` by :user:`Guillaume Lemaitre `. - |Feature| :func:`metrics.roc_auc_score` now supports micro-averaging (`average="micro"`) for the One-vs-Rest multiclass case (`multi_class="ovr"`). :pr:`24338` by :user:`Arturo Amor `. - |Enhancement| Adds an `"auto"` option to `eps` in :func:`metrics.logloss`. This option will automatically set the `eps` value depending on the data type of `y_pred`. In addition, the default value of `eps` is changed from `1e-15` to the new `"auto"` option. :pr:`24354` by :user:`Safiuddin Khaja ` and :user:`gsiisg `. - |Fix| Allows `csr_matrix` as input for parameter: `y_true` of the :func:`metrics.label_ranking_average_precision_score` metric. :pr:`23442` by :user:`Sean Atukorala ` - |Fix| :func:`metrics.ndcg_score` will now trigger a warning when the `y_true` value contains a negative value. Users may still use negative values, but the result may not be between 0 and 1. Starting in v1.4, passing in negative values for `y_true` will raise an error. :pr:`22710` by :user:`Conroy Trinh ` and :pr:`23461` by :user:`Meekail Zain `. - |Fix| :func:`metrics.log_loss` with `eps=0` now returns a correct value of 0 or `np.inf` instead of `nan` for predictions at the boundaries (0 or 1). It also accepts integer input. :pr:`24365` by :user:`Christian Lorentzen `. - |API| The parameter `sum_over_features` of :func:`metrics.pairwise.manhattan_distances` is deprecated and will be removed in 1.4. :pr:`24630` by :user:`Rushil Desai `. :mod:`sklearn.model_selection` .............................. - |Feature| Added the class :class:`model_selection.LearningCurveDisplay` that allows to make easy plotting of learning curves obtained by the function :func:`model_selection.learning_curve`. :pr:`24084` by :user:`Guillaume Lemaitre `. - |Fix| For all `SearchCV` classes and scipy >= 1.10, rank corresponding to a nan score is correctly set to the maximum possible rank, rather than `np.iinfo(np.int32).min`. :pr:`24141` by :user:`Loïc Estève `. - |Fix| In both :class:`model_selection.HalvingGridSearchCV` and :class:`model_selection.HalvingRandomSearchCV` parameter combinations with a NaN score now share the lowest rank. :pr:`24539` by :user:`Tim Head `. - |Fix| For :class:`model_selection.GridSearchCV` and :class:`model_selection.RandomizedSearchCV` ranks corresponding to nan scores will all be set to the maximum possible rank. :pr:`24543` by :user:`Guillaume Lemaitre `. :mod:`sklearn.multioutput` .......................... - |Feature| Added boolean `verbose` flag to classes: :class:`multioutput.ClassifierChain` and :class:`multioutput.RegressorChain`. :pr:`23977` by :user:`Eric Fiegel `, :user:`Chiara Marmo `, :user:`Lucy Liu `, and :user:`Guillaume Lemaitre `. :mod:`sklearn.naive_bayes` .......................... - |Feature| Add methods `predict_joint_log_proba` to all naive Bayes classifiers. :pr:`23683` by :user:`Andrey Melnik `. - |Enhancement| A new parameter `force_alpha` was added to :class:`naive_bayes.BernoulliNB`, :class:`naive_bayes.ComplementNB`, :class:`naive_bayes.CategoricalNB`, and :class:`naive_bayes.MultinomialNB`, allowing user to set parameter alpha to a very small number, greater or equal 0, which was earlier automatically changed to `1e-10` instead. :pr:`16747` by :user:`arka204`, :pr:`18805` by :user:`hongshaoyang`, :pr:`22269` by :user:`Meekail Zain `. :mod:`sklearn.neighbors` ........................ - |Feature| Adds new function :func:`neighbors.sort_graph_by_row_values` to sort a CSR sparse graph such that each row is stored with increasing values. This is useful to improve efficiency when using precomputed sparse distance matrices in a variety of estimators and avoid an `EfficiencyWarning`. :pr:`23139` by `Tom Dupre la Tour`_. - |Efficiency| :class:`neighbors.NearestCentroid` is faster and requires less memory as it better leverages CPUs' caches to compute predictions. :pr:`24645` by :user:`Olivier Grisel `. - |Enhancement| :class:`neighbors.KernelDensity` bandwidth parameter now accepts definition using Scott's and Silverman's estimation methods. :pr:`10468` by :user:`Ruben ` and :pr:`22993` by :user:`Jovan Stojanovic `. - |Enhancement| :class:`neighbors.NeighborsBase` now accepts Minkowski semi-metric (i.e. when :math:`0 < p < 1` for `metric="minkowski"`) for `algorithm="auto"` or `algorithm="brute"`. :pr:`24750` by :user:`Rudresh Veerkhare ` - |Fix| :class:`neighbors.NearestCentroid` now raises an informative error message at fit-time instead of failing with a low-level error message at predict-time. :pr:`23874` by :user:`Juan Gomez <2357juan>`. - |Fix| Set `n_jobs=None` by default (instead of `1`) for :class:`neighbors.KNeighborsTransformer` and :class:`neighbors.RadiusNeighborsTransformer`. :pr:`24075` by :user:`Valentin Laurent `. - |Enhancement| :class:`neighbors.LocalOutlierFactor` now preserves dtype for `numpy.float32` inputs. :pr:`22665` by :user:`Julien Jerphanion `. :mod:`sklearn.neural_network` ............................. - |Fix| :class:`neural_network.MLPClassifier` and :class:`neural_network.MLPRegressor` always expose the parameters `best_loss_`, `validation_scores_`, and `best_validation_score_`. `best_loss_` is set to `None` when `early_stopping=True`, while `validation_scores_` and `best_validation_score_` are set to `None` when `early_stopping=False`. :pr:`24683` by :user:`Guillaume Lemaitre `. :mod:`sklearn.pipeline` ....................... - |Enhancement| :meth:`pipeline.FeatureUnion.get_feature_names_out` can now be used when one of the transformers in the :class:`pipeline.FeatureUnion` is `"passthrough"`. :pr:`24058` by :user:`Diederik Perdok ` - |Enhancement| The :class:`pipeline.FeatureUnion` class now has a `named_transformers` attribute for accessing transformers by name. :pr:`20331` by :user:`Christopher Flynn `. :mod:`sklearn.preprocessing` ............................ - |Enhancement| :class:`preprocessing.FunctionTransformer` will always try to set `n_features_in_` and `feature_names_in_` regardless of the `validate` parameter. :pr:`23993` by `Thomas Fan`_. - |Fix| :class:`preprocessing.LabelEncoder` correctly encodes NaNs in `transform`. :pr:`22629` by `Thomas Fan`_. - |API| The `sparse` parameter of :class:`preprocessing.OneHotEncoder` is now deprecated and will be removed in version 1.4. Use `sparse_output` instead. :pr:`24412` by :user:`Rushil Desai `. :mod:`sklearn.svm` .................. - |API| The `class_weight_` attribute is now deprecated for :class:`svm.NuSVR`, :class:`svm.SVR`, :class:`svm.OneClassSVM`. :pr:`22898` by :user:`Meekail Zain `. :mod:`sklearn.tree` ................... - |Enhancement| :func:`tree.plot_tree`, :func:`tree.export_graphviz` now uses a lower case `x[i]` to represent feature `i`. :pr:`23480` by `Thomas Fan`_. :mod:`sklearn.utils` .................... - |Feature| A new module exposes development tools to discover estimators (i.e. :func:`utils.discovery.all_estimators`), displays (i.e. :func:`utils.discovery.all_displays`) and functions (i.e. :func:`utils.discovery.all_functions`) in scikit-learn. :pr:`21469` by :user:`Guillaume Lemaitre `. - |Enhancement| :func:`utils.extmath.randomized_svd` now accepts an argument, `lapack_svd_driver`, to specify the lapack driver used in the internal deterministic SVD used by the randomized SVD algorithm. :pr:`20617` by :user:`Srinath Kailasa ` - |Enhancement| :func:`utils.validation.column_or_1d` now accepts a `dtype` parameter to specific `y`'s dtype. :pr:`22629` by `Thomas Fan`_. - |Enhancement| :func:`utils.extmath.cartesian` now accepts arrays with different `dtype` and will cast the ouptut to the most permissive `dtype`. :pr:`25067` by :user:`Guillaume Lemaitre `. - |Fix| :func:`utils.multiclass.type_of_target` now properly handles sparse matrices. :pr:`14862` by :user:`Léonard Binet `. - |Fix| HTML representation no longer errors when an estimator class is a value in `get_params`. :pr:`24512` by `Thomas Fan`_. - |Fix| :func:`utils.estimator_checks.check_estimator` now takes into account the `requires_positive_X` tag correctly. :pr:`24667` by `Thomas Fan`_. - |Fix| :func:`utils.check_array` now supports Pandas Series with `pd.NA` by raising a better error message or returning a compatible `ndarray`. :pr:`25080` by `Thomas Fan`_. - |API| The extra keyword parameters of :func:`utils.extmath.density` are deprecated and will be removed in 1.4. :pr:`24523` by :user:`Mia Bajic `. Code and Documentation Contributors ----------------------------------- Thanks to everyone who has contributed to the maintenance and improvement of the project since version 1.1, including: 2357juan, 3lLobo, Adam J. Stewart, Adam Li, Aditya Anulekh, Adrin Jalali, Aiko, Akshita Prasanth, Alessandro Miola, Alex, Alexandr, Alexandre Perez-Lebel, aman kumar, Amit Bera, Andreas Grivas, Andreas Mueller, angela-maennel, Aniket Shirsat, Antony Lee, anupam, Apostolos Tsetoglou, Aravindh R, Arturo Amor, Ashwin Mathur, avm19, b0rxington, Badr MOUFAD, Bardiya Ak, Bartłomiej Gońda, BdeGraaff, Benjamin Carter, berkecanrizai, Bernd Fritzke, Bhoomika, Biswaroop Mitra, Brandon TH Chen, Brett Cannon, Bsh, carlo, Carlos Ramos Carreño, ceh, chalulu, Charles Zablit, Chiara Marmo, Christian Lorentzen, Christian Ritter, christianwaldmann, Christine P. Chai, Claudio Salvatore Arcidiacono, Clément Verrier, Daniela Fernandes, DanielGaerber, darioka, Darren Nguyen, David Gilbertson, David Poznik, Dev Khant, Dhanshree Arora, Diadochokinetic, diederikwp, Dimitri Papadopoulos Orfanos, drewhogg, Duarte OC, Dwight Lindquist, Eden Brekke, Edoardo Abati, Eleanore Denies, EliaSchiavon, ErmolaevPA, Fabrizio Damicelli, fcharras, Flynn, francesco-tuveri, Franck Charras, ftorres16, Gael Varoquaux, Geevarghese George, GeorgiaMayDay, Gianr Lazz, Gleb Levitski, Glòria Macià Muñoz, Guillaume Lemaitre, Guillem García Subies, Guitared, Hansin Ahuja, Hao Chun Chang, Harsh Agrawal, harshit5674, hasan-yaman, henrymooresc, Henry Sorsky, Hristo Vrigazov, htsedebenham, humahn, i-aki-y, Iglesys, Iliya Zhechev, Irene, ivanllt, Ivan Sedykh, jakirkham, Jason G, Jérémie du Boisberranger, Jiten Sidhpura, jkarolczak, João David, John Koumentis, John P, johnthagen, Jordan Fleming, Joshua Choo Yun Keat, Jovan Stojanovic, Juan Carlos Alfaro Jiménez, juanfe88, Juan Felipe Arias, Julien Jerphanion, Kanishk Sachdev, Kanissh, Kendall, Kenneth Prabakaran, kernc, Kevin Roice, Kian Eliasi, Kilian Lieret, Kirandevraj, Kraig, krishna kumar, krishna vamsi, Kshitij Kapadni, Kshitij Mathur, Lauren Burke, Léonard Binet, lingyi1110, Lisa Casino, Loic Esteve, Luciano Mantovani, Lucy Liu, Maascha, Madhura Jayaratne, madinak, Maksym, Malte S. Kurz, Mansi Agrawal, Marco Edward Gorelli, Marco Wurps, Maren Westermann, Maria Telenczuk, martin-kokos, mathurinm, mauroantonioserrano, Maxi Marufo, Maxim Smolskiy, Maxwell, Meekail Zain, Mehgarg, mehmetcanakbay, Mia Bajić, Michael Flaks, Michael Hornstein, Michel de Ruiter, Michelle Paradis, Misa Ogura, Moritz Wilksch, mrastgoo, Naipawat Poolsawat, Naoise Holohan, Nass, Nathan Jacobi, Nguyễn Văn Diễn, Nihal Thukarama Rao, Nikita Jare, nima10khodaveisi, Nima Sarajpoor, nitinramvelraj, Nwanna-Joseph, Nymark Kho, o-holman, Olivier Grisel, Olle Lukowski, Omar Hassoun, Omar Salman, osman tamer, Oyindamola Olatunji, Paulo Sergio Soares, Petar Mlinarić, Peter Jansson, Peter Steinbach, Philipp Jung, Piet Brömmel, priyam kakati, puhuk, Rachel Freeland, Rachit Keerti Das, Rahil Parikh, Ravi Makhija, Rehan Guha, Reshama Shaikh, Richard Klima, Rob Crockett, Robert Hommes, Robert Juergens, Robin Lenz, Rocco Meli, Roman4oo, Ross Barnowski, Rowan Mankoo, Rudresh Veerkhare, Rushil Desai, Sabri Monaf Sabri, Safikh, Safiuddin Khaja, Salahuddin, Sam Adam Day, Sandra Yojana Meneses, Sandro Ephrem, Sangam, SangamSwadik, SarahRemus, SavkoMax, Sean Atukorala, sec65, SELEE, seljaks, Shane, shellyfung, Shinsuke Mori, Shoaib Khan, Shrankhla Srivastava, Shuangchi He, Simon, Srinath Kailasa, Stefanie Molin, stellalin7, Stéphane Collot, Steve Schmerler, Sven Stehle, TheDevPanda, the-syd-sre, Thomas Bonald, Thomas J. Fan, Ti-Ion, Tim Head, Timofei Kornev, toastedyeast, Tobias Pitters, Tom Dupré la Tour, Tom Mathews, Tom McTiernan, tspeng, Tyler Egashira, Valentin Laurent, Varun Jain, Vera Komeyer, Vicente Reyes-Puerta, Vincent M, Vishal, wattai, wchathura, WEN Hao, x110, Xiao Yuan, Xunius, yanhong-zhao-ef, Z Adil Khwaja