scikit-learn / scikit-learn Public

37 Pull requests merged by 17 people

• MAINT Bump versions to match oldest-supported-numpy

  #22674 merged Mar 4, 2022

• ENH Introduce `PairwiseDistancesRadiusNeighborhood`

  #22320 merged Mar 4, 2022

• DOC Fixes visualization docs and runs code to plot

  #22579 merged Mar 4, 2022

• DOC: Update notebook-style example plot_dbscan

  #22568 merged Mar 3, 2022

• MNT Replace PriorityHeap with cpp heap methods in trees

  #22630 merged Mar 3, 2022

• MAINT Remove deprecated exceptions for 1.1

  #22659 merged Mar 3, 2022

• MNT More cleaning for 1.1

  #22661 merged Mar 3, 2022

• DOC Add minimal reproducer guide for scikit-learn

  #22401 merged Mar 3, 2022

• DOC Add directives about writing bibliographic references.

  #22656 merged Mar 2, 2022

• DOC Sphinx warnings.

  #22660 merged Mar 2, 2022

• FIX Reduces memory usage of `PCA.transform`

  #22553 merged Mar 2, 2022

• Use :doi: and :arxiv: directives for references in the documentation (Issue #21088)

  #22603 merged Mar 2, 2022

• BLD Places newest highlights first in subsection

  #22620 merged Mar 2, 2022

• FIX better handle limit cases in normalized_mutual_info_score

  #22635 merged Mar 2, 2022

• CI Adds upstream sync to check manifest

  #22641 merged Mar 2, 2022

• MNT remaining clean up for 1.1

  #22643 merged Mar 2, 2022

• DOC use notebook-style for plot_faces_decomposition

  #22452 merged Mar 2, 2022

• FIX Make Ridge*CV warn about rescaling alphas with scaling

  #22585 merged Mar 2, 2022

• MNT Clean fixes and compat for old versions of our dependencies

  #22642 merged Mar 2, 2022

• CI Use a job specific cache key for the datasets in CircleCI

  #22650 merged Mar 2, 2022

• DOC fix matplotlib warning for plot_ward_structured_vs_unstructured.py

  #22631 merged Mar 2, 2022

• DOC Fix matplotlib warning for plot_iris_dataset.py

  #22640 merged Mar 2, 2022

• MAINT pypy compat for sklearn.show_versions

  #22649 merged Mar 2, 2022

• MAINT update black config

  #22647 merged Mar 1, 2022

• BLD Exclude SECURITY.md from MANIFEST

  #22638 merged Feb 28, 2022

• FIX Avoid side effects importing pip/setuptool in show_version

  #22621 merged Feb 28, 2022

• CI Label PRs only when it is opened

  #22637 merged Feb 28, 2022

• MNT Drops Python 3.7 in CI, wheel building, and docs

  #22617 merged Feb 28, 2022

• CLN Small refactor of cost complexity pruning using _TREE_UNDEFINED

  #22351 merged Feb 28, 2022

• ENH Isomap supports radius-based neighbors

  #19794 merged Feb 28, 2022

• DOC Add abstract methods to _BaseDiscreteNB

  #22596 merged Feb 28, 2022

• ENH Bayesian regression model `dtype` consistency

  #22525 merged Feb 27, 2022

• DOC Fix axes3d matplotlib deprecation warning

  #22622 merged Feb 27, 2022

• DOC Fix typo in decomposition.TruncatedSVD

  #22624 merged Feb 27, 2022

• DOC Ensures that linnerud passes numpydoc validation

  #22484 merged Feb 27, 2022

• MNT Add validation for parameter alphas in LassoCV

  #22305 merged Feb 26, 2022

• MNT Fixes pypy load_files error in test

  #22619 merged Feb 26, 2022

26 Pull requests opened by 5 people

• FIX Improves nan support in LabelEncoder

  #22629 opened Feb 28, 2022

• Fix Axes3D matplotlib deprecation warning on example plot_partial_dep…

  #22654 opened Mar 2, 2022

• TST Adapt `test_neighbors.py` to test implementations on 32bit datasets

  #22663 opened Mar 3, 2022

• TST Adapt `test_lof.py` to test implementations on 32bit datasets

  #22665 opened Mar 3, 2022

• TST Adapt `test_pairwise.py` to test implementations on 32bit datasets

  #22666 opened Mar 3, 2022

• TST Adapt `test_affinity_propagation.py` to test implementations on 32bit datasets.

  #22667 opened Mar 3, 2022

• TST Adapt `test_optics.py` to test implementations on 32bit datasets

  #22668 opened Mar 3, 2022

• TST Adapt `test_spectral.py` to test implementations on 32bit datasets

  #22669 opened Mar 3, 2022

• TST Adapt `test_label_propagation.py` to test implementations on 32bit …

  #22670 opened Mar 3, 2022

• TST Adapt `test_birch.py` to test implementations on 32bit datasets

  #22671 opened Mar 3, 2022

• TST Adapt `test_mean_shift.py` to test implementations on 32bit datasets

  #22672 opened Mar 3, 2022

• TST Adapt `test_isomap.py` to test implementations on 32bit datasets

  #22673 opened Mar 3, 2022

• TST Adapt `test_t_sne.py` to test implementations on 32bit datasets

  #22675 opened Mar 3, 2022

• TST Adapt `test_locally_linear.py` to test implementations on 32bit datasets

  #22676 opened Mar 3, 2022

• TST Adapt `test_mutual_info.py` to test implementations on 32bit datasets

  #22677 opened Mar 3, 2022

• DOC use notebook-style and fix matplotlib warning for plot_compare_methods

  #22679 opened Mar 3, 2022

• DOC Standardizing the documentation for `alpha` in `Ridge` and `Lasso`

  #22684 opened Mar 3, 2022

• ENH Adds `metric_params` parameter to pass to metrics in `TSNE`

  #22685 opened Mar 3, 2022

• FIX Make `KNeighborsRegressor.predict` compatible with array-likes

  #22687 opened Mar 4, 2022

• TST Add minimal setup to be able to run test suite on float32

  #22690 opened Mar 4, 2022

• BLD Monkeypatch windows build to stablize build

  #22693 opened Mar 4, 2022

• ENH Adds feature_names_out for most of kernel_approximation

  #22694 opened Mar 4, 2022

• ENH Adds feature_names_out to stacking estimators

  #22695 opened Mar 4, 2022

• FIX Correct `LDA` `fac` multiplier

  #22696 opened Mar 4, 2022

• ENH Add feature_names_out to voting estimators

  #22697 opened Mar 5, 2022

• DOC Fixes VotingClassifier.transform docstring

  #22698 opened Mar 5, 2022

23 Issues closed by 7 people

• Unexpected output from Random Forest Classifer

  #22692 closed Mar 4, 2022

• Visualizations examples in documentation don't work

  #22563 closed Mar 4, 2022

• Document how to craft a minimal reproducer for scikit-learn

  #21945 closed Mar 3, 2022

• Enforce references to use `:doi:` and `arxiv` directives.

  #22627 closed Mar 2, 2022

• Deferred calculation of expensive fitted attributes

  #22552 closed Mar 2, 2022

• Randomized PCA.transform uses a lot of RAM

  #11102 closed Mar 2, 2022

• SimpleImputer does not support pd.NA

  #22655 closed Mar 2, 2022

• DOC: use :doi: and :arxiv: directives for references

  #21088 closed Mar 2, 2022

• Release highlights icon in sphinx is misordered

  #21994 closed Mar 2, 2022

• Odd (incorrect) behavior with normalized_mutual_info_score

  #13836 closed Mar 2, 2022

• RidgeCV replacement for normalize=True gives different results

  #22540 closed Mar 2, 2022

• joblib development version breaks `sklearn.show_versions()`

  #22614 closed Feb 28, 2022

• Don't run the labeller twice on a same PR

  #22636 closed Feb 28, 2022

• Adding Documentation about creating Virtual Environment for Contribution using Anaconda

  #14756 closed Feb 28, 2022

• No inverse function for grid_to_graph(..., mask=M) ?

  #14578 closed Feb 27, 2022

• Add feature encoding algorithms (Fisher vector, VLAD, and BoW)

  #14382 closed Feb 27, 2022

• check_regressors_train failure on master with the latest release on conda

  #14106 closed Feb 27, 2022

• add lightgbm to the CI

  #14210 closed Feb 27, 2022

• PyPy 7.1.1: open(file_path), argument must be a string without NUL characters

  #14174 closed Feb 27, 2022

• PowerTransformer(‘box-cox’) got different results with same data in different platform(windows, linux)

  #14089 closed Feb 27, 2022

• Extending tensorflow model with BaseEstimator throws a thread lock error

  #11410 closed Feb 27, 2022

• Different Probability scores in different systems

  #11940 closed Feb 27, 2022

• Transformers need to be explained better in tutorial material

  #3210 closed Feb 27, 2022

13 Issues opened by 12 people

• Installing scipy-wheels-nightly using pip shows error

  #22699 opened Mar 5, 2022

• Include entire range in `check_scalar` error message

  #22691 opened Mar 4, 2022

• kMeans stopped working with numpy 1.22.2

  #22689 opened Mar 4, 2022

• KNeighborsRegressor with a callable weights stopped working with numpy 1.22.2

  #22683 opened Mar 3, 2022

• Estimator check for dtype preservation for regressors

  #22682 opened Mar 3, 2022

• TST Add option to run tests on 32bit data

  #22680 opened Mar 3, 2022

• GridSearchCV does not return trained estimator for each split vs cross_validate which does

  #22678 opened Mar 3, 2022

• Add NaN handling in sklearn.preprocessing.KBinsDiscretizer

  #22664 opened Mar 3, 2022

• multiple-output regression is inaccurate for many targets

  #22645 opened Mar 1, 2022

• Add `as_frame` to `sklearn.datasets.fetch_20newsgroups`

  #22644 opened Mar 1, 2022

• LabelEncoder not transforming nans as expected.

  #22628 opened Feb 27, 2022

• Add an average callable parameter to sklearn.neighbors.KNeighborsRegressor

  #22626 opened Feb 27, 2022

• Add zero_division None or np.nan

  #22625 opened Feb 27, 2022

113 Unresolved conversations

Sometimes conversations happen on old items that aren’t yet closed. Here is a list of all the Issues and Pull Requests with unresolved conversations.

• Base sample-prop implementation and docs (alternative to #21284)

  #22083 commented on Mar 4, 2022 • 44 new comments

• ENH Adds infrequent categories to OneHotEncoder

  #16018 commented on Mar 4, 2022 • 31 new comments

• [MRG] Refactor MiniBatchDictionaryLearning and add stopping criterion

  #18975 commented on Mar 4, 2022 • 24 new comments

• Unable to successfully install scikit-learn development version with MacBookPro with chip Apple M1 Pro

  #22581 commented on Mar 4, 2022 • 23 new comments

• [MRG] separate penalty factors for GLM regressions

  #22485 commented on Mar 4, 2022 • 17 new comments

• MAINT 32bit datasets support for `PairwiseDistancesReduction`

  #22590 commented on Mar 4, 2022 • 14 new comments

• FEA Bisecting K-Means

  #20031 commented on Feb 28, 2022 • 12 new comments

• [MRG] Online implementation of non-negative matrix factorization

  #16948 commented on Mar 3, 2022 • 10 new comments

• MNT Clean deprecation of dtype='numeric' + array of strings in check_array for 1.1

  #22589 commented on Mar 4, 2022 • 7 new comments

• Fix Axes3D matplotlib deprecation warning in examples

  #22586 commented on Mar 2, 2022 • 5 new comments

• FIX Support extra class_weights in compute_class_weight

  #22595 commented on Mar 4, 2022 • 5 new comments

• ndcg_score fails for negative scores

  #17639 commented on Mar 3, 2022 • 3 new comments

• Clarify group order in GroupKFold and LeaveOneGroupOut

  #22582 commented on Mar 2, 2022 • 3 new comments

• add which solver was used in sklearn Ridge regression : add clf.get_params(solver)

  #22341 commented on Feb 27, 2022 • 2 new comments

• Add friedman's H statistic

  #22383 commented on Mar 1, 2022 • 2 new comments

• Use the function check_scalar for parameters validation

  #21927 commented on Mar 3, 2022 • 2 new comments

• [MRG] Added option to return raw predictions from cross_validate

  #18390 commented on Mar 1, 2022 • 2 new comments

• FIX #18692, updates the doc

  #18781 commented on Mar 4, 2022 • 2 new comments

• DOC Improve the mathematical description of Logistic Regression

  #22382 commented on Mar 3, 2022 • 2 new comments

• [WIP] Add `eigh` solver to FastICA

  #22527 commented on Mar 5, 2022 • 2 new comments

• ENH Adds Array API support to LinearDiscriminantAnalysis

  #22554 commented on Mar 1, 2022 • 2 new comments

• DOC updated to notebook style for grid_search_text_feature_extraction.py

  #22558 commented on Feb 27, 2022 • 2 new comments

• ENH add Naive Bayes Metaestimator `ColumnwiseNB` (aka "GeneralNB")

  #22574 commented on Mar 4, 2022 • 2 new comments

• base_estimator vs estimator for meta-estimators

  #9104 commented on Feb 26, 2022 • 1 new comment

• Use a stable sorting algorithm when selecting the `max_features` in TfidfVectorizer

  #21446 commented on Feb 26, 2022 • 1 new comment

• All kernel regression methods should accept precomputed Gram matrices

  #8445 commented on Feb 27, 2022 • 1 new comment

• Support inverse_transform in ColumnTransformer

  #11463 commented on Feb 27, 2022 • 1 new comment

• Bisecting K-means

  #14214 commented on Feb 28, 2022 • 1 new comment

• Add Penalty factors for each coefficient in enet (similar to the R's glmnet library)

  #11566 commented on Feb 28, 2022 • 1 new comment

• max_iter limit argument inconsistencies between solvers.

  #18205 commented on Mar 1, 2022 • 1 new comment

• Change name of Triage Team

  #21993 commented on Mar 1, 2022 • 1 new comment

• feature: verbose flag in cluster.DBscan

  #14947 commented on Mar 1, 2022 • 1 new comment

• The default value of some parameters make little sense

  #13570 commented on Mar 3, 2022 • 1 new comment

• Ensure that functions's docstrings pass numpydoc validation

  #21350 commented on Mar 3, 2022 • 1 new comment

• Fix notebook-style examples

  #22406 commented on Mar 4, 2022 • 1 new comment

• Warning the user of bad default values, starting by dbscan.eps

  #14942 commented on Mar 3, 2022 • 1 new comment

• solved bug in _initialize precisions_cholesky_

  #16910 commented on Feb 28, 2022 • 1 new comment

• [MRG] Paired haversine distances

  #17258 commented on Mar 4, 2022 • 1 new comment

• Draft implementation of quantile methods

  #19754 commented on Feb 28, 2022 • 1 new comment

• ENH Preserving dtype for np.float32 in LatentDirichletAllocation

  #22113 commented on Mar 4, 2022 • 1 new comment

• ENH Optimise `decomposition.FastICA.fit` memory footprint and runtime

  #22268 commented on Mar 4, 2022 • 1 new comment

• Kernel Ridge Classifier

  #22417 commented on Mar 1, 2022 • 1 new comment

• ENH Make sure OneHotEncoder returns unique feature names

  #22506 commented on Feb 26, 2022 • 1 new comment

• FEA Add positive and negative likelihood ratios to metrics

  #22518 commented on Mar 4, 2022 • 1 new comment

• DOC Link directly developer docs in the navbar

  #22550 commented on Mar 1, 2022 • 1 new comment

• DOC: use notebook-style for plot_svm_anova.py

  #22555 commented on Feb 26, 2022 • 1 new comment

• `RidgeCV` `cv_values_` are for preprocessed data: centered and scaled by sample weights.

  #13998 commented on Feb 27, 2022 • 0 new comments

• [RFC] API / objects for out of core fitting, partial_fit pipelines

  #4834 commented on Feb 27, 2022 • 0 new comments

• Invariance testing for partial_fit

  #3896 commented on Feb 27, 2022 • 0 new comments

• Working with float32 data

  #5776 commented on Feb 27, 2022 • 0 new comments

• Inconsistent output shapes for Regressors

  #5058 commented on Feb 27, 2022 • 0 new comments

• RFC binary operators on estimators

  #7608 commented on Feb 27, 2022 • 0 new comments

• Establish global error state like np.seterr

  #4660 commented on Feb 27, 2022 • 0 new comments

• Found a case where isotonic regression takes non-polynomial time

  #8063 commented on Feb 27, 2022 • 0 new comments

• Classification scorers requiring categorical predictions should not be applied to regressors

  #9267 commented on Feb 27, 2022 • 0 new comments

• Relationship between different Group-based (and perhaps general) CV splitters should be better explained

  #9385 commented on Feb 27, 2022 • 0 new comments

• retain original indexes

  #8238 commented on Feb 27, 2022 • 0 new comments

• New feature: correlation matrix generation

  #9900 commented on Feb 27, 2022 • 0 new comments

• Proposal to decouple classifier output classes from the y vector with a classes= parameter in __init__

  #1183 commented on Feb 27, 2022 • 0 new comments

• Inconsistent density estimator sampling APIs

  #10539 commented on Feb 27, 2022 • 0 new comments

• Pandas Interoperability section in User Guide

  #10453 commented on Feb 27, 2022 • 0 new comments

• Remove Hardcoded Thresholds in Repr

  #11897 commented on Feb 27, 2022 • 0 new comments

• Joblib makes callstacks impossible to decipher

  #12322 commented on Feb 27, 2022 • 0 new comments

• R-like summary of LogisticRegression

  #12575 commented on Feb 27, 2022 • 0 new comments

• Univariate feature selection methods should allow passing kwargs to score_func callable

  #12468 commented on Feb 27, 2022 • 0 new comments

• LDA covariance shrinkage 'auto' bug

  #12662 commented on Feb 27, 2022 • 0 new comments

• Add post-prune option for decision tree as well as randomForest and GBDT?

  #4630 commented on Feb 27, 2022 • 0 new comments

• [regression] --record {plist} records a directory

  #12681 commented on Feb 27, 2022 • 0 new comments

• repr of column transformer unhelpful

  #13770 commented on Feb 27, 2022 • 0 new comments

• Support for Pandas DataFrames in kernels & heterogeneous kernel

  #13668 commented on Feb 27, 2022 • 0 new comments

• Add common test for .transform on different sparse formats

  #14176 commented on Feb 27, 2022 • 0 new comments

• Library should not be printing to stdout without verbose option

  #6929 commented on Feb 27, 2022 • 0 new comments

• Error while using n_jobs=-1 with uwsgi

  #12922 commented on Feb 27, 2022 • 0 new comments

• Using pre-fitted GradientBoostingRegressor init estimator

  #12947 commented on Feb 27, 2022 • 0 new comments

• RFC Common internal optimizers API

  #14248 commented on Feb 27, 2022 • 0 new comments

• Add Matrix Profile algorithm to scikit-learn

  #14011 commented on Feb 27, 2022 • 0 new comments

• Stale scikit-learn datasets can throw pickling errors

  #14328 commented on Feb 27, 2022 • 0 new comments

• LassoLarsIC is missing Xy from its fit method signature which is present in parent parent method.

  #14145 commented on Feb 27, 2022 • 0 new comments

• test_sample_order_invariance in common metrics tests applied to threshold metrics

  #12308 commented on Feb 27, 2022 • 0 new comments

• OpenML: avoid warning when all versions are inactive

  #14829 commented on Feb 28, 2022 • 0 new comments

• [RANSACRegressor] Strange behaviour with rectangular regions

  #22560 commented on Feb 28, 2022 • 0 new comments

• RFC Principled metrics for scoring and calibration of supervised learning

  #21718 commented on Feb 28, 2022 • 0 new comments

• MAINT: unified treatment of MACHINE_EPSILON and np.finfo in the code base

  #5609 commented on Feb 28, 2022 • 0 new comments

• Pipeline requires both fit and transform method to be available instead of only fit_transform

  #16710 commented on Mar 1, 2022 • 0 new comments

• Improving non-interactive Usage

  #13012 commented on Mar 3, 2022 • 0 new comments

• Feature addition: Pipeline comparison

  #9631 commented on Mar 3, 2022 • 0 new comments

• Make AIF360 default bias checker and mitigator in scikit-learn

  #14181 commented on Mar 3, 2022 • 0 new comments

• data leak in GBDT due to warm start

  #14034 commented on Mar 4, 2022 • 0 new comments

• [QST] Updating indices creation for split

  #14036 commented on Mar 4, 2022 • 0 new comments

• `prange` in trees

  #14037 commented on Mar 4, 2022 • 0 new comments

• TREE: public function/method to return a new estimator with a new ccp_alpha

  #14038 commented on Mar 4, 2022 • 0 new comments

• TREE: warm_start with ccp_alpha

  #14039 commented on Mar 4, 2022 • 0 new comments

• var_smoothing parameter of GaussianNB isn't documented nor tested

  #14054 commented on Mar 4, 2022 • 0 new comments

• `PairwiseDistancesReductions` roadmap

  #22587 commented on Mar 4, 2022 • 0 new comments

• Implement `get_feature_names_out` for all estimators

  #21308 commented on Mar 5, 2022 • 0 new comments

• [WIP] Common test for equivalence between sparse and dense matrices.

  #13246 commented on Mar 4, 2022 • 0 new comments

• [MRG] Monotonic Contraints for Tree-based models

  #13649 commented on Mar 4, 2022 • 0 new comments

• Solve Gaussian Mixture Model sample/kmeans++-based initialisation merge conflicts #11101

  #20408 commented on Mar 2, 2022 • 0 new comments

• ENH improve ARFF parser using pandas

  #21938 commented on Mar 2, 2022 • 0 new comments

• ENH Adds encoded_missing_value to OrdinalEncoder

  #21988 commented on Mar 3, 2022 • 0 new comments

• FIX make coef_ in PLS estimator consistent with linear models

  #22016 commented on Mar 2, 2022 • 0 new comments

• ENH Adds get_feature_names_out for AdditiveChi2Sampler

  #22137 commented on Mar 3, 2022 • 0 new comments

• [MRG] Adding variable `force_alpha` to classes in naive_bayes.py

  #22269 commented on Feb 28, 2022 • 0 new comments

• [MRG] Update gridsearch example for multimetric scoring.

  #22279 commented on Mar 4, 2022 • 0 new comments

• ENH Implement SVD algorithm in MDS

  #22330 commented on Feb 27, 2022 • 0 new comments

• ENH Allow `SelectFromModel`'s `max_features` to accept callables

  #22356 commented on Mar 4, 2022 • 0 new comments

• API Deprecate max_feature=`auto` for tree classes

  #22476 commented on Mar 2, 2022 • 0 new comments

• FIX Remove validation from __init__ and set_params for ColumnTransformer

  #22537 commented on Mar 3, 2022 • 0 new comments

• DOC: Update notebook-style example plot_affinity_propagation

  #22559 commented on Feb 28, 2022 • 0 new comments

• ENH Feature: calculate normed stress (Stress-1) in `manifold.MDS`

  #22562 commented on Feb 27, 2022 • 0 new comments

• FIX Fixes OneVsOneClassifier.predict for Estimators with only predict_proba

  #22604 commented on Mar 2, 2022 • 0 new comments

• [WIP] Include HDBSCAN as a sub-module for `sklearn.cluster`

  #22616 commented on Mar 5, 2022 • 0 new comments

• (WIP) ENH Add zero_division = none or np.nan

  #22618 commented on Feb 26, 2022 • 0 new comments