-
Notifications
You must be signed in to change notification settings - Fork 81
/
Copy pathv0.22.html
1440 lines (1405 loc) · 212 KB
/
v0.22.html
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
<!DOCTYPE html>
<!--[if IE 8]><html class="no-js lt-ie9" lang="en" > <![endif]-->
<!--[if gt IE 8]><!--> <html class="no-js" lang="en" > <!--<![endif]-->
<head>
<meta charset="utf-8">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<meta name="Description" content="scikit-learn: machine learning in Python">
<title>Version 0.22.2.post1 — scikit-learn 0.22.2 documentation</title>
<link rel="canonical" href="https://scikit-learn.org/stable/whats_new/v0.22.html" />
<link rel="shortcut icon" href="../_static/favicon.ico"/>
<link rel="stylesheet" href="../_static/css/vendor/bootstrap.min.css" type="text/css" />
<link rel="stylesheet" href="../_static/gallery.css" type="text/css" />
<link rel="stylesheet" href="../_static/css/theme.css" type="text/css" />
<script id="documentation_options" data-url_root="../" src="../_static/documentation_options.js"></script>
<script src="../_static/jquery.js"></script>
</head>
<body>
<nav id="navbar" class="sk-docs-navbar navbar navbar-expand-md navbar-light bg-light py-0">
<div class="container-fluid sk-docs-container px-0">
<a class="navbar-brand py-0" href="../index.html">
<img
class="sk-brand-img"
src="../_static/scikit-learn-logo-small.png"
alt="logo"/>
</a>
<button
id="sk-navbar-toggler"
class="navbar-toggler"
type="button"
data-toggle="collapse"
data-target="#navbarSupportedContent"
aria-controls="navbarSupportedContent"
aria-expanded="false"
aria-label="Toggle navigation"
>
<span class="navbar-toggler-icon"></span>
</button>
<div class="sk-navbar-collapse collapse navbar-collapse" id="navbarSupportedContent">
<ul class="navbar-nav mr-auto">
<li class="nav-item">
<a class="sk-nav-link nav-link" href="../install.html">Install</a>
</li>
<li class="nav-item">
<a class="sk-nav-link nav-link" href="../user_guide.html">User Guide</a>
</li>
<li class="nav-item">
<a class="sk-nav-link nav-link" href="../modules/classes.html">API</a>
</li>
<li class="nav-item">
<a class="sk-nav-link nav-link" href="../auto_examples/index.html">Examples</a>
</li>
<li class="nav-item">
<a class="sk-nav-link nav-link nav-more-item-mobile-items" href="../getting_started.html">Getting Started</a>
</li>
<li class="nav-item">
<a class="sk-nav-link nav-link nav-more-item-mobile-items" href="../tutorial/index.html">Tutorial</a>
</li>
<li class="nav-item">
<a class="sk-nav-link nav-link nav-more-item-mobile-items" href="../glossary.html">Glossary</a>
</li>
<li class="nav-item">
<a class="sk-nav-link nav-link nav-more-item-mobile-items" href="../developers/index.html">Development</a>
</li>
<li class="nav-item">
<a class="sk-nav-link nav-link nav-more-item-mobile-items" href="../faq.html">FAQ</a>
</li>
<li class="nav-item">
<a class="sk-nav-link nav-link nav-more-item-mobile-items" href="../related_projects.html">Related packages</a>
</li>
<li class="nav-item">
<a class="sk-nav-link nav-link nav-more-item-mobile-items" href="../roadmap.html">Roadmap</a>
</li>
<li class="nav-item">
<a class="sk-nav-link nav-link nav-more-item-mobile-items" href="../about.html">About us</a>
</li>
<li class="nav-item">
<a class="sk-nav-link nav-link nav-more-item-mobile-items" href="https://github.com/scikit-learn/scikit-learn">GitHub</a>
</li>
<li class="nav-item">
<a class="sk-nav-link nav-link nav-more-item-mobile-items" href="https://scikit-learn.org/dev/versions.html">Other Versions</a>
</li>
<li class="nav-item dropdown nav-more-item-dropdown">
<a class="sk-nav-link nav-link dropdown-toggle" href="#" id="navbarDropdown" role="button" data-toggle="dropdown" aria-haspopup="true" aria-expanded="false">More</a>
<div class="dropdown-menu" aria-labelledby="navbarDropdown">
<a class="sk-nav-dropdown-item dropdown-item" href="../getting_started.html">Getting Started</a>
<a class="sk-nav-dropdown-item dropdown-item" href="../tutorial/index.html">Tutorial</a>
<a class="sk-nav-dropdown-item dropdown-item" href="../glossary.html">Glossary</a>
<a class="sk-nav-dropdown-item dropdown-item" href="../developers/index.html">Development</a>
<a class="sk-nav-dropdown-item dropdown-item" href="../faq.html">FAQ</a>
<a class="sk-nav-dropdown-item dropdown-item" href="../related_projects.html">Related packages</a>
<a class="sk-nav-dropdown-item dropdown-item" href="../roadmap.html">Roadmap</a>
<a class="sk-nav-dropdown-item dropdown-item" href="../about.html">About us</a>
<a class="sk-nav-dropdown-item dropdown-item" href="https://github.com/scikit-learn/scikit-learn">GitHub</a>
<a class="sk-nav-dropdown-item dropdown-item" href="https://scikit-learn.org/dev/versions.html">Other Versions</a>
</div>
</li>
</ul>
<div id="searchbox" role="search">
<div class="searchformwrapper">
<form class="search" action="../search.html" method="get">
<input class="sk-search-text-input" type="text" name="q" aria-labelledby="searchlabel" />
<input class="sk-search-text-btn" type="submit" value="Go" />
</form>
</div>
</div>
</div>
</div>
</nav>
<div class="d-flex" id="sk-doc-wrapper">
<input type="checkbox" name="sk-toggle-checkbox" id="sk-toggle-checkbox">
<label id="sk-sidemenu-toggle" class="sk-btn-toggle-toc btn sk-btn-primary" for="sk-toggle-checkbox">Toggle Menu</label>
<div id="sk-sidebar-wrapper" class="border-right">
<div class="sk-sidebar-toc-wrapper">
<div class="sk-sidebar-toc-logo">
<a href="../index.html">
<img
class="sk-brand-img"
src="../_static/scikit-learn-logo-small.png"
alt="logo"/>
</a>
</div>
<div class="btn-group w-100 mb-2" role="group" aria-label="rellinks">
<a href="../whats_new.html" role="button" class="btn sk-btn-rellink py-1" sk-rellink-tooltip="Release History">Prev</a><a href="../whats_new.html" role="button" class="btn sk-btn-rellink py-1" sk-rellink-tooltip="Release History">Up</a>
<a href="v0.21.html" role="button" class="btn sk-btn-rellink py-1" sk-rellink-tooltip="Version 0.21.3">Next</a>
</div>
<div class="alert alert-danger p-1 mb-2" role="alert">
<p class="text-center mb-0">
<strong>scikit-learn 0.22.2</strong><br/>
<a href="http://scikit-learn.org/dev/versions.html">Other versions</a>
</p>
</div>
<div class="alert alert-warning p-1 mb-2" role="alert">
<p class="text-center mb-0">
Please <a class="font-weight-bold" href="../about.html#citing-scikit-learn"><string>cite us</string></a> if you use the software.
</p>
</div>
<div class="sk-sidebar-toc">
<ul>
<li><a class="reference internal" href="#">Version 0.22.2.post1</a><ul>
<li><a class="reference internal" href="#changelog">Changelog</a><ul>
<li><a class="reference internal" href="#sklearn-impute"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.impute</span></code></a></li>
<li><a class="reference internal" href="#sklearn-metrics"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.metrics</span></code></a></li>
<li><a class="reference internal" href="#sklearn-neighbors"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.neighbors</span></code></a></li>
</ul>
</li>
</ul>
</li>
<li><a class="reference internal" href="#version-0-22-1">Version 0.22.1</a><ul>
<li><a class="reference internal" href="#id1">Changelog</a><ul>
<li><a class="reference internal" href="#sklearn-cluster"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.cluster</span></code></a></li>
<li><a class="reference internal" href="#sklearn-inspection"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.inspection</span></code></a></li>
<li><a class="reference internal" href="#id2"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.metrics</span></code></a></li>
<li><a class="reference internal" href="#sklearn-model-selection"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.model_selection</span></code></a></li>
<li><a class="reference internal" href="#sklearn-naive-bayes"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.naive_bayes</span></code></a></li>
<li><a class="reference internal" href="#sklearn-preprocessing"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.preprocessing</span></code></a></li>
<li><a class="reference internal" href="#sklearn-semi-supervised"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.semi_supervised</span></code></a></li>
<li><a class="reference internal" href="#sklearn-utils"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.utils</span></code></a></li>
</ul>
</li>
</ul>
</li>
<li><a class="reference internal" href="#version-0-22-0">Version 0.22.0</a><ul>
<li><a class="reference internal" href="#legend-for-changelogs">Legend for changelogs</a></li>
<li><a class="reference internal" href="#website-update">Website update</a></li>
<li><a class="reference internal" href="#clear-definition-of-the-public-api">Clear definition of the public API</a></li>
<li><a class="reference internal" href="#deprecations-using-futurewarning-from-now-on">Deprecations: using <code class="docutils literal notranslate"><span class="pre">FutureWarning</span></code> from now on</a></li>
<li><a class="reference internal" href="#changed-models">Changed models</a></li>
<li><a class="reference internal" href="#id4">Changelog</a><ul>
<li><a class="reference internal" href="#sklearn-base"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.base</span></code></a></li>
<li><a class="reference internal" href="#sklearn-calibration"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.calibration</span></code></a></li>
<li><a class="reference internal" href="#id5"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.cluster</span></code></a></li>
<li><a class="reference internal" href="#sklearn-compose"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.compose</span></code></a></li>
<li><a class="reference internal" href="#sklearn-cross-decomposition"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.cross_decomposition</span></code></a></li>
<li><a class="reference internal" href="#sklearn-datasets"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.datasets</span></code></a></li>
<li><a class="reference internal" href="#sklearn-decomposition"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.decomposition</span></code></a></li>
<li><a class="reference internal" href="#sklearn-dummy"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.dummy</span></code></a></li>
<li><a class="reference internal" href="#sklearn-ensemble"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.ensemble</span></code></a></li>
<li><a class="reference internal" href="#sklearn-feature-extraction"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.feature_extraction</span></code></a></li>
<li><a class="reference internal" href="#sklearn-feature-selection"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.feature_selection</span></code></a></li>
<li><a class="reference internal" href="#sklearn-gaussian-process"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.gaussian_process</span></code></a></li>
<li><a class="reference internal" href="#id6"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.impute</span></code></a></li>
<li><a class="reference internal" href="#id7"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.inspection</span></code></a></li>
<li><a class="reference internal" href="#sklearn-kernel-approximation"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.kernel_approximation</span></code></a></li>
<li><a class="reference internal" href="#sklearn-linear-model"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.linear_model</span></code></a></li>
<li><a class="reference internal" href="#sklearn-manifold"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.manifold</span></code></a></li>
<li><a class="reference internal" href="#id8"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.metrics</span></code></a></li>
<li><a class="reference internal" href="#id9"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.model_selection</span></code></a></li>
<li><a class="reference internal" href="#sklearn-multioutput"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.multioutput</span></code></a></li>
<li><a class="reference internal" href="#id10"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.naive_bayes</span></code></a></li>
<li><a class="reference internal" href="#id11"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.neighbors</span></code></a></li>
<li><a class="reference internal" href="#sklearn-neural-network"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.neural_network</span></code></a></li>
<li><a class="reference internal" href="#sklearn-pipeline"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.pipeline</span></code></a></li>
<li><a class="reference internal" href="#id12"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.preprocessing</span></code></a></li>
<li><a class="reference internal" href="#id13"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.model_selection</span></code></a></li>
<li><a class="reference internal" href="#sklearn-svm"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.svm</span></code></a></li>
<li><a class="reference internal" href="#sklearn-tree"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.tree</span></code></a></li>
<li><a class="reference internal" href="#id14"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.utils</span></code></a></li>
<li><a class="reference internal" href="#sklearn-isotonic"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.isotonic</span></code></a></li>
<li><a class="reference internal" href="#miscellaneous">Miscellaneous</a></li>
</ul>
</li>
<li><a class="reference internal" href="#changes-to-estimator-checks">Changes to estimator checks</a></li>
<li><a class="reference internal" href="#code-and-documentation-contributors">Code and Documentation Contributors</a></li>
</ul>
</li>
</ul>
</div>
</div>
</div>
<div id="sk-page-content-wrapper">
<div class="sk-page-content container-fluid body px-md-3" role="main">
<div class="section" id="version-0-22-2-post1">
<span id="changes-0-22-2"></span><h1>Version 0.22.2.post1<a class="headerlink" href="#version-0-22-2-post1" title="Permalink to this headline">¶</a></h1>
<p><strong>March 3 2020</strong></p>
<p>The 0.22.2.post1 release includes a packaging fix for the source distribution
but the content of the packages is otherwise identical to the content of the
wheels with the 0.22.2 version (without the .post1 suffix). Both contain the
following changes.</p>
<div class="section" id="changelog">
<h2>Changelog<a class="headerlink" href="#changelog" title="Permalink to this headline">¶</a></h2>
<div class="section" id="sklearn-impute">
<h3><a class="reference internal" href="../modules/classes.html#module-sklearn.impute" title="sklearn.impute"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.impute</span></code></a><a class="headerlink" href="#sklearn-impute" title="Permalink to this headline">¶</a></h3>
<ul class="simple">
<li><p><span class="raw-html"><span class="badge badge-info">Efficiency</span></span> Reduce <a class="reference internal" href="../modules/generated/sklearn.impute.KNNImputer.html#sklearn.impute.KNNImputer" title="sklearn.impute.KNNImputer"><code class="xref py py-func docutils literal notranslate"><span class="pre">impute.KNNImputer</span></code></a> asymptotic memory usage by
chunking pairwise distance computation.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/16397">#16397</a> by <a class="reference external" href="https://joelnothman.com/">Joel Nothman</a>.</p></li>
</ul>
</div>
<div class="section" id="sklearn-metrics">
<h3><a class="reference internal" href="../modules/classes.html#module-sklearn.metrics" title="sklearn.metrics"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.metrics</span></code></a><a class="headerlink" href="#sklearn-metrics" title="Permalink to this headline">¶</a></h3>
<ul class="simple">
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> Fixed a bug in <a class="reference internal" href="../modules/generated/sklearn.metrics.plot_roc_curve.html#sklearn.metrics.plot_roc_curve" title="sklearn.metrics.plot_roc_curve"><code class="xref py py-func docutils literal notranslate"><span class="pre">metrics.plot_roc_curve</span></code></a> where
the name of the estimator was passed in the <a class="reference internal" href="../modules/generated/sklearn.metrics.RocCurveDisplay.html#sklearn.metrics.RocCurveDisplay" title="sklearn.metrics.RocCurveDisplay"><code class="xref py py-class docutils literal notranslate"><span class="pre">metrics.RocCurveDisplay</span></code></a>
instead of the parameter <code class="docutils literal notranslate"><span class="pre">name</span></code>. It results in a different plot when calling
<a class="reference internal" href="../modules/generated/sklearn.metrics.RocCurveDisplay.html#sklearn.metrics.RocCurveDisplay.plot" title="sklearn.metrics.RocCurveDisplay.plot"><code class="xref py py-meth docutils literal notranslate"><span class="pre">metrics.RocCurveDisplay.plot</span></code></a> for the subsequent times.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/16500">#16500</a> by <a class="reference external" href="https://github.com/glemaitre">Guillaume Lemaitre</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> Fixed a bug in <a class="reference internal" href="../modules/generated/sklearn.metrics.plot_precision_recall_curve.html#sklearn.metrics.plot_precision_recall_curve" title="sklearn.metrics.plot_precision_recall_curve"><code class="xref py py-func docutils literal notranslate"><span class="pre">metrics.plot_precision_recall_curve</span></code></a> where the
name of the estimator was passed in the
<a class="reference internal" href="../modules/generated/sklearn.metrics.PrecisionRecallDisplay.html#sklearn.metrics.PrecisionRecallDisplay" title="sklearn.metrics.PrecisionRecallDisplay"><code class="xref py py-class docutils literal notranslate"><span class="pre">metrics.PrecisionRecallDisplay</span></code></a> instead of the parameter <code class="docutils literal notranslate"><span class="pre">name</span></code>. It
results in a different plot when calling
<a class="reference internal" href="../modules/generated/sklearn.metrics.PrecisionRecallDisplay.html#sklearn.metrics.PrecisionRecallDisplay.plot" title="sklearn.metrics.PrecisionRecallDisplay.plot"><code class="xref py py-meth docutils literal notranslate"><span class="pre">metrics.PrecisionRecallDisplay.plot</span></code></a> for the subsequent times.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/16505">#16505</a> by <a class="reference external" href="https://github.com/glemaitre">Guillaume Lemaitre</a>.</p></li>
</ul>
</div>
<div class="section" id="sklearn-neighbors">
<h3><a class="reference internal" href="../modules/classes.html#module-sklearn.neighbors" title="sklearn.neighbors"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.neighbors</span></code></a><a class="headerlink" href="#sklearn-neighbors" title="Permalink to this headline">¶</a></h3>
<ul class="simple">
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> Fix a bug which converted a list of arrays into a 2-D object
array instead of a 1-D array containing NumPy arrays. This bug
was affecting <a class="reference internal" href="../modules/generated/sklearn.neighbors.NearestNeighbors.html#sklearn.neighbors.NearestNeighbors.radius_neighbors" title="sklearn.neighbors.NearestNeighbors.radius_neighbors"><code class="xref py py-meth docutils literal notranslate"><span class="pre">neighbors.NearestNeighbors.radius_neighbors</span></code></a>.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/16076">#16076</a> by <a class="reference external" href="https://github.com/glemaitre">Guillaume Lemaitre</a> and
<a class="reference external" href="https://github.com/alexshacked">Alex Shacked</a>.</p></li>
</ul>
</div>
</div>
</div>
<div class="section" id="version-0-22-1">
<span id="changes-0-22-1"></span><h1>Version 0.22.1<a class="headerlink" href="#version-0-22-1" title="Permalink to this headline">¶</a></h1>
<p><strong>January 2 2020</strong></p>
<p>This is a bug-fix release to primarily resolve some packaging issues in version
0.22.0. It also includes minor documentation improvements and some bug fixes.</p>
<div class="section" id="id1">
<h2>Changelog<a class="headerlink" href="#id1" title="Permalink to this headline">¶</a></h2>
<div class="section" id="sklearn-cluster">
<h3><a class="reference internal" href="../modules/classes.html#module-sklearn.cluster" title="sklearn.cluster"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.cluster</span></code></a><a class="headerlink" href="#sklearn-cluster" title="Permalink to this headline">¶</a></h3>
<ul class="simple">
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> <a class="reference internal" href="../modules/generated/sklearn.cluster.KMeans.html#sklearn.cluster.KMeans" title="sklearn.cluster.KMeans"><code class="xref py py-class docutils literal notranslate"><span class="pre">cluster.KMeans</span></code></a> with <code class="docutils literal notranslate"><span class="pre">algorithm="elkan"</span></code> now uses the same
stopping criterion as with the default <code class="docutils literal notranslate"><span class="pre">algorithm="full"</span></code>. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/15930">#15930</a> by
<a class="reference external" href="https://github.com/inder128">@inder128</a>.</p></li>
</ul>
</div>
<div class="section" id="sklearn-inspection">
<h3><a class="reference internal" href="../modules/classes.html#module-sklearn.inspection" title="sklearn.inspection"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.inspection</span></code></a><a class="headerlink" href="#sklearn-inspection" title="Permalink to this headline">¶</a></h3>
<ul class="simple">
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> <a class="reference internal" href="../modules/generated/sklearn.inspection.permutation_importance.html#sklearn.inspection.permutation_importance" title="sklearn.inspection.permutation_importance"><code class="xref py py-func docutils literal notranslate"><span class="pre">inspection.permutation_importance</span></code></a> will return the same
<code class="docutils literal notranslate"><span class="pre">importances</span></code> when a <code class="docutils literal notranslate"><span class="pre">random_state</span></code> is given for both <code class="docutils literal notranslate"><span class="pre">n_jobs=1</span></code> or
<code class="docutils literal notranslate"><span class="pre">n_jobs>1</span></code> both with shared memory backends (thread-safety) and
isolated memory, process-based backends.
Also avoid casting the data as object dtype and avoid read-only error
on large dataframes with <code class="docutils literal notranslate"><span class="pre">n_jobs>1</span></code> as reported in <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/issues/15810">#15810</a>.
Follow-up of <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/15898">#15898</a> by <a class="reference external" href="https://github.com/shivamgargsya">Shivam Gargsya</a>.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/15933">#15933</a> by <a class="reference external" href="https://github.com/glemaitre">Guillaume Lemaitre</a> and <a class="reference external" href="https://twitter.com/ogrisel">Olivier Grisel</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> <a class="reference internal" href="../modules/generated/sklearn.inspection.plot_partial_dependence.html#sklearn.inspection.plot_partial_dependence" title="sklearn.inspection.plot_partial_dependence"><code class="xref py py-func docutils literal notranslate"><span class="pre">inspection.plot_partial_dependence</span></code></a> and
<a class="reference internal" href="../modules/generated/sklearn.inspection.PartialDependenceDisplay.html#sklearn.inspection.PartialDependenceDisplay.plot" title="sklearn.inspection.PartialDependenceDisplay.plot"><code class="xref py py-meth docutils literal notranslate"><span class="pre">inspection.PartialDependenceDisplay.plot</span></code></a> now consistently checks
the number of axes passed in. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/15760">#15760</a> by <a class="reference external" href="https://github.com/thomasjpfan">Thomas Fan</a>.</p></li>
</ul>
</div>
<div class="section" id="id2">
<h3><a class="reference internal" href="../modules/classes.html#module-sklearn.metrics" title="sklearn.metrics"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.metrics</span></code></a><a class="headerlink" href="#id2" title="Permalink to this headline">¶</a></h3>
<ul class="simple">
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> <a class="reference internal" href="../modules/generated/sklearn.metrics.plot_confusion_matrix.html#sklearn.metrics.plot_confusion_matrix" title="sklearn.metrics.plot_confusion_matrix"><code class="xref py py-func docutils literal notranslate"><span class="pre">metrics.plot_confusion_matrix</span></code></a> now raises error when <code class="docutils literal notranslate"><span class="pre">normalize</span></code>
is invalid. Previously, it runs fine with no normalization.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/15888">#15888</a> by <a class="reference external" href="https://github.com/qinhanmin2014">Hanmin Qin</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> <a class="reference internal" href="../modules/generated/sklearn.metrics.plot_confusion_matrix.html#sklearn.metrics.plot_confusion_matrix" title="sklearn.metrics.plot_confusion_matrix"><code class="xref py py-func docutils literal notranslate"><span class="pre">metrics.plot_confusion_matrix</span></code></a> now colors the label color
correctly to maximize contrast with its background. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/15936">#15936</a> by
<a class="reference external" href="https://github.com/thomasjpfan">Thomas Fan</a> and <a class="reference external" href="https://github.com/DizietAsahi">@DizietAsahi</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> <a class="reference internal" href="../modules/generated/sklearn.metrics.classification_report.html#sklearn.metrics.classification_report" title="sklearn.metrics.classification_report"><code class="xref py py-func docutils literal notranslate"><span class="pre">metrics.classification_report</span></code></a> does no longer ignore the
value of the <code class="docutils literal notranslate"><span class="pre">zero_division</span></code> keyword argument. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/15879">#15879</a>
by <a class="reference external" href="https://github.com/Bibyutatsu">Bibhash Chandra Mitra</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> Fixed a bug in <a class="reference internal" href="../modules/generated/sklearn.metrics.plot_confusion_matrix.html#sklearn.metrics.plot_confusion_matrix" title="sklearn.metrics.plot_confusion_matrix"><code class="xref py py-func docutils literal notranslate"><span class="pre">metrics.plot_confusion_matrix</span></code></a> to correctly
pass the <code class="docutils literal notranslate"><span class="pre">values_format</span></code> parameter to the <code class="xref py py-class docutils literal notranslate"><span class="pre">ConfusionMatrixDisplay</span></code>
plot() call. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/15937">#15937</a> by <a class="reference external" href="https://github.com/blynotes">Stephen Blystone</a>.</p></li>
</ul>
</div>
<div class="section" id="sklearn-model-selection">
<h3><a class="reference internal" href="../modules/classes.html#module-sklearn.model_selection" title="sklearn.model_selection"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.model_selection</span></code></a><a class="headerlink" href="#sklearn-model-selection" title="Permalink to this headline">¶</a></h3>
<ul class="simple">
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> <a class="reference internal" href="../modules/generated/sklearn.model_selection.GridSearchCV.html#sklearn.model_selection.GridSearchCV" title="sklearn.model_selection.GridSearchCV"><code class="xref py py-class docutils literal notranslate"><span class="pre">model_selection.GridSearchCV</span></code></a> and
<a class="reference internal" href="../modules/generated/sklearn.model_selection.RandomizedSearchCV.html#sklearn.model_selection.RandomizedSearchCV" title="sklearn.model_selection.RandomizedSearchCV"><code class="xref py py-class docutils literal notranslate"><span class="pre">model_selection.RandomizedSearchCV</span></code></a> accept scalar values provided in
<code class="docutils literal notranslate"><span class="pre">fit_params</span></code>. Change in 0.22 was breaking backward compatibility.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/15863">#15863</a> by <a class="reference external" href="https://github.com/adrinjalali">Adrin Jalali</a> and
<a class="reference external" href="https://github.com/glemaitre">Guillaume Lemaitre</a>.</p></li>
</ul>
</div>
<div class="section" id="sklearn-naive-bayes">
<h3><a class="reference internal" href="../modules/classes.html#module-sklearn.naive_bayes" title="sklearn.naive_bayes"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.naive_bayes</span></code></a><a class="headerlink" href="#sklearn-naive-bayes" title="Permalink to this headline">¶</a></h3>
<ul class="simple">
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> Removed <code class="docutils literal notranslate"><span class="pre">abstractmethod</span></code> decorator for the method <code class="docutils literal notranslate"><span class="pre">_check_X</span></code> in
<code class="xref py py-class docutils literal notranslate"><span class="pre">naive_bayes.BaseNB</span></code> that could break downstream projects inheriting
from this deprecated public base class. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/15996">#15996</a> by
<a class="reference external" href="https://github.com/bsipocz">Brigitta Sipőcz</a>.</p></li>
</ul>
</div>
<div class="section" id="sklearn-preprocessing">
<h3><a class="reference internal" href="../modules/classes.html#module-sklearn.preprocessing" title="sklearn.preprocessing"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.preprocessing</span></code></a><a class="headerlink" href="#sklearn-preprocessing" title="Permalink to this headline">¶</a></h3>
<ul class="simple">
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> <a class="reference internal" href="../modules/generated/sklearn.preprocessing.QuantileTransformer.html#sklearn.preprocessing.QuantileTransformer" title="sklearn.preprocessing.QuantileTransformer"><code class="xref py py-class docutils literal notranslate"><span class="pre">preprocessing.QuantileTransformer</span></code></a> now guarantees the
<code class="docutils literal notranslate"><span class="pre">quantiles_</span></code> attribute to be completely sorted in non-decreasing manner.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/15751">#15751</a> by <a class="reference external" href="https://github.com/tirthasheshpatel">Tirth Patel</a>.</p></li>
</ul>
</div>
<div class="section" id="sklearn-semi-supervised">
<h3><a class="reference internal" href="../modules/classes.html#module-sklearn.semi_supervised" title="sklearn.semi_supervised"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.semi_supervised</span></code></a><a class="headerlink" href="#sklearn-semi-supervised" title="Permalink to this headline">¶</a></h3>
<ul class="simple">
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> <a class="reference internal" href="../modules/generated/sklearn.semi_supervised.LabelPropagation.html#sklearn.semi_supervised.LabelPropagation" title="sklearn.semi_supervised.LabelPropagation"><code class="xref py py-class docutils literal notranslate"><span class="pre">semi_supervised.LabelPropagation</span></code></a> and
<a class="reference internal" href="../modules/generated/sklearn.semi_supervised.LabelSpreading.html#sklearn.semi_supervised.LabelSpreading" title="sklearn.semi_supervised.LabelSpreading"><code class="xref py py-class docutils literal notranslate"><span class="pre">semi_supervised.LabelSpreading</span></code></a> now allow callable kernel function to
return sparse weight matrix.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/15868">#15868</a> by <a class="reference external" href="https://github.com/nik-sm">Niklas Smedemark-Margulies</a>.</p></li>
</ul>
</div>
<div class="section" id="sklearn-utils">
<h3><a class="reference internal" href="../modules/classes.html#module-sklearn.utils" title="sklearn.utils"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.utils</span></code></a><a class="headerlink" href="#sklearn-utils" title="Permalink to this headline">¶</a></h3>
<ul class="simple">
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> <a class="reference internal" href="../modules/generated/sklearn.utils.check_array.html#sklearn.utils.check_array" title="sklearn.utils.check_array"><code class="xref py py-func docutils literal notranslate"><span class="pre">utils.check_array</span></code></a> now correctly converts pandas DataFrame with
boolean columns to floats. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/15797">#15797</a> by <a class="reference external" href="https://github.com/thomasjpfan">Thomas Fan</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> <code class="xref py py-func docutils literal notranslate"><span class="pre">utils.check_is_fitted</span></code> accepts back an explicit <code class="docutils literal notranslate"><span class="pre">attributes</span></code>
argument to check for specific attributes as explicit markers of a fitted
estimator. When no explicit <code class="docutils literal notranslate"><span class="pre">attributes</span></code> are provided, only the attributes
that end with a underscore and do not start with double underscore are used
as “fitted” markers. The <code class="docutils literal notranslate"><span class="pre">all_or_any</span></code> argument is also no longer
deprecated. This change is made to restore some backward compatibility with
the behavior of this utility in version 0.21. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/15947">#15947</a> by <a class="reference external" href="https://github.com/thomasjpfan">Thomas Fan</a>.</p></li>
</ul>
</div>
</div>
</div>
<div class="section" id="version-0-22-0">
<span id="changes-0-22"></span><h1>Version 0.22.0<a class="headerlink" href="#version-0-22-0" title="Permalink to this headline">¶</a></h1>
<p><strong>December 3 2019</strong></p>
<p>For a short description of the main highlights of the release, please
refer to
<a class="reference internal" href="../auto_examples/release_highlights/plot_release_highlights_0_22_0.html#sphx-glr-auto-examples-release-highlights-plot-release-highlights-0-22-0-py"><span class="std std-ref">Release Highlights for scikit-learn 0.22</span></a>.</p>
<div class="section" id="legend-for-changelogs">
<h2>Legend for changelogs<a class="headerlink" href="#legend-for-changelogs" title="Permalink to this headline">¶</a></h2>
<ul class="simple">
<li><p><span class="raw-html"><span class="badge badge-success">Major Feature</span></span> : something big that you couldn’t do before.</p></li>
<li><p><span class="raw-html"><span class="badge badge-success">Feature</span></span> : something that you couldn’t do before.</p></li>
<li><p><span class="raw-html"><span class="badge badge-info">Efficiency</span></span> : an existing feature now may not require as much computation or
memory.</p></li>
<li><p><span class="raw-html"><span class="badge badge-info">Enhancement</span></span> : a miscellaneous minor improvement.</p></li>
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> : something that previously didn’t work as documentated – or according
to reasonable expectations – should now work.</p></li>
<li><p><span class="raw-html"><span class="badge badge-warning">API Change</span></span> : you will need to change your code to have the same effect in the
future; or a feature will be removed in the future.</p></li>
</ul>
</div>
<div class="section" id="website-update">
<h2>Website update<a class="headerlink" href="#website-update" title="Permalink to this headline">¶</a></h2>
<p><a class="reference external" href="https://scikit-learn.org/">Our website</a> was revamped and given a fresh
new look. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/14849">#14849</a> by <a class="reference external" href="https://github.com/thomasjpfan">Thomas Fan</a>.</p>
</div>
<div class="section" id="clear-definition-of-the-public-api">
<h2>Clear definition of the public API<a class="headerlink" href="#clear-definition-of-the-public-api" title="Permalink to this headline">¶</a></h2>
<p>Scikit-learn has a public API, and a private API.</p>
<p>We do our best not to break the public API, and to only introduce
backward-compatible changes that do not require any user action. However, in
cases where that’s not possible, any change to the public API is subject to
a deprecation cycle of two minor versions. The private API isn’t publicly
documented and isn’t subject to any deprecation cycle, so users should not
rely on its stability.</p>
<p>A function or object is public if it is documented in the <a class="reference external" href="https://scikit-learn.org/dev/modules/classes.html">API Reference</a> and if it can be
imported with an import path without leading underscores. For example
<code class="docutils literal notranslate"><span class="pre">sklearn.pipeline.make_pipeline</span></code> is public, while
<code class="docutils literal notranslate"><span class="pre">sklearn.pipeline._name_estimators</span></code> is private.
<code class="docutils literal notranslate"><span class="pre">sklearn.ensemble._gb.BaseEnsemble</span></code> is private too because the whole <code class="docutils literal notranslate"><span class="pre">_gb</span></code>
module is private.</p>
<p>Up to 0.22, some tools were de-facto public (no leading underscore), while
they should have been private in the first place. In version 0.22, these
tools have been made properly private, and the public API space has been
cleaned. In addition, importing from most sub-modules is now deprecated: you
should for example use <code class="docutils literal notranslate"><span class="pre">from</span> <span class="pre">sklearn.cluster</span> <span class="pre">import</span> <span class="pre">Birch</span></code> instead of
<code class="docutils literal notranslate"><span class="pre">from</span> <span class="pre">sklearn.cluster.birch</span> <span class="pre">import</span> <span class="pre">Birch</span></code> (in practice, <code class="docutils literal notranslate"><span class="pre">birch.py</span></code> has
been moved to <code class="docutils literal notranslate"><span class="pre">_birch.py</span></code>).</p>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>All the tools in the public API should be documented in the <a class="reference external" href="https://scikit-learn.org/dev/modules/classes.html">API
Reference</a>. If you
find a public tool (without leading underscore) that isn’t in the API
reference, that means it should either be private or documented. Please
let us know by opening an issue!</p>
</div>
<p>This work was tracked in <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/issues/9250">issue 9250</a> and <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/issues/12927">issue
12927</a>.</p>
</div>
<div class="section" id="deprecations-using-futurewarning-from-now-on">
<h2>Deprecations: using <code class="docutils literal notranslate"><span class="pre">FutureWarning</span></code> from now on<a class="headerlink" href="#deprecations-using-futurewarning-from-now-on" title="Permalink to this headline">¶</a></h2>
<p>When deprecating a feature, previous versions of scikit-learn used to raise
a <code class="docutils literal notranslate"><span class="pre">DeprecationWarning</span></code>. Since the <code class="docutils literal notranslate"><span class="pre">DeprecationWarnings</span></code> aren’t shown by
default by Python, scikit-learn needed to resort to a custom warning filter
to always show the warnings. That filter would sometimes interfere
with users custom warning filters.</p>
<p>Starting from version 0.22, scikit-learn will show <code class="docutils literal notranslate"><span class="pre">FutureWarnings</span></code> for
deprecations, <a class="reference external" href="https://docs.python.org/3/library/exceptions.html#FutureWarning">as recommended by the Python documentation</a>.
<code class="docutils literal notranslate"><span class="pre">FutureWarnings</span></code> are always shown by default by Python, so the custom
filter has been removed and scikit-learn no longer hinders with user
filters. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/15080">#15080</a> by <a class="reference external" href="https://github.com/NicolasHug">Nicolas Hug</a>.</p>
</div>
<div class="section" id="changed-models">
<h2>Changed models<a class="headerlink" href="#changed-models" title="Permalink to this headline">¶</a></h2>
<p>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.</p>
<ul class="simple">
<li><p><a class="reference internal" href="../modules/generated/sklearn.cluster.KMeans.html#sklearn.cluster.KMeans" title="sklearn.cluster.KMeans"><code class="xref py py-class docutils literal notranslate"><span class="pre">cluster.KMeans</span></code></a> when <code class="docutils literal notranslate"><span class="pre">n_jobs=1</span></code>. <span class="raw-html"><span class="badge badge-danger">Fix</span></span> </p></li>
<li><p><a class="reference internal" href="../modules/generated/sklearn.decomposition.SparseCoder.html#sklearn.decomposition.SparseCoder" title="sklearn.decomposition.SparseCoder"><code class="xref py py-class docutils literal notranslate"><span class="pre">decomposition.SparseCoder</span></code></a>,
<a class="reference internal" href="../modules/generated/sklearn.decomposition.DictionaryLearning.html#sklearn.decomposition.DictionaryLearning" title="sklearn.decomposition.DictionaryLearning"><code class="xref py py-class docutils literal notranslate"><span class="pre">decomposition.DictionaryLearning</span></code></a>, and
<a class="reference internal" href="../modules/generated/sklearn.decomposition.MiniBatchDictionaryLearning.html#sklearn.decomposition.MiniBatchDictionaryLearning" title="sklearn.decomposition.MiniBatchDictionaryLearning"><code class="xref py py-class docutils literal notranslate"><span class="pre">decomposition.MiniBatchDictionaryLearning</span></code></a> <span class="raw-html"><span class="badge badge-danger">Fix</span></span> </p></li>
<li><p><a class="reference internal" href="../modules/generated/sklearn.decomposition.SparseCoder.html#sklearn.decomposition.SparseCoder" title="sklearn.decomposition.SparseCoder"><code class="xref py py-class docutils literal notranslate"><span class="pre">decomposition.SparseCoder</span></code></a> with <code class="docutils literal notranslate"><span class="pre">algorithm='lasso_lars'</span></code> <span class="raw-html"><span class="badge badge-danger">Fix</span></span> </p></li>
<li><p><a class="reference internal" href="../modules/generated/sklearn.decomposition.SparsePCA.html#sklearn.decomposition.SparsePCA" title="sklearn.decomposition.SparsePCA"><code class="xref py py-class docutils literal notranslate"><span class="pre">decomposition.SparsePCA</span></code></a> where <code class="docutils literal notranslate"><span class="pre">normalize_components</span></code> has no effect
due to deprecation.</p></li>
<li><p><a class="reference internal" href="../modules/generated/sklearn.ensemble.HistGradientBoostingClassifier.html#sklearn.ensemble.HistGradientBoostingClassifier" title="sklearn.ensemble.HistGradientBoostingClassifier"><code class="xref py py-class docutils literal notranslate"><span class="pre">ensemble.HistGradientBoostingClassifier</span></code></a> and
<a class="reference internal" href="../modules/generated/sklearn.ensemble.HistGradientBoostingRegressor.html#sklearn.ensemble.HistGradientBoostingRegressor" title="sklearn.ensemble.HistGradientBoostingRegressor"><code class="xref py py-class docutils literal notranslate"><span class="pre">ensemble.HistGradientBoostingRegressor</span></code></a> <span class="raw-html"><span class="badge badge-danger">Fix</span></span> , <span class="raw-html"><span class="badge badge-success">Feature</span></span> ,
<span class="raw-html"><span class="badge badge-info">Enhancement</span></span> .</p></li>
<li><p><a class="reference internal" href="../modules/generated/sklearn.impute.IterativeImputer.html#sklearn.impute.IterativeImputer" title="sklearn.impute.IterativeImputer"><code class="xref py py-class docutils literal notranslate"><span class="pre">impute.IterativeImputer</span></code></a> when <code class="docutils literal notranslate"><span class="pre">X</span></code> has features with no missing
values. <span class="raw-html"><span class="badge badge-success">Feature</span></span> </p></li>
<li><p><a class="reference internal" href="../modules/generated/sklearn.linear_model.Ridge.html#sklearn.linear_model.Ridge" title="sklearn.linear_model.Ridge"><code class="xref py py-class docutils literal notranslate"><span class="pre">linear_model.Ridge</span></code></a> when <code class="docutils literal notranslate"><span class="pre">X</span></code> is sparse. <span class="raw-html"><span class="badge badge-danger">Fix</span></span> </p></li>
<li><p><a class="reference internal" href="../modules/generated/sklearn.model_selection.StratifiedKFold.html#sklearn.model_selection.StratifiedKFold" title="sklearn.model_selection.StratifiedKFold"><code class="xref py py-class docutils literal notranslate"><span class="pre">model_selection.StratifiedKFold</span></code></a> and any use of <code class="docutils literal notranslate"><span class="pre">cv=int</span></code> with a
classifier. <span class="raw-html"><span class="badge badge-danger">Fix</span></span> </p></li>
<li><p><a class="reference internal" href="../modules/generated/sklearn.cross_decomposition.CCA.html#sklearn.cross_decomposition.CCA" title="sklearn.cross_decomposition.CCA"><code class="xref py py-class docutils literal notranslate"><span class="pre">cross_decomposition.CCA</span></code></a> when using scipy >= 1.3 <span class="raw-html"><span class="badge badge-danger">Fix</span></span> </p></li>
</ul>
<p>Details are listed in the changelog below.</p>
<p>(While we are trying to better inform users by providing this information, we
cannot assure that this list is complete.)</p>
</div>
<div class="section" id="id4">
<h2>Changelog<a class="headerlink" href="#id4" title="Permalink to this headline">¶</a></h2>
<div class="section" id="sklearn-base">
<h3><a class="reference internal" href="../modules/classes.html#module-sklearn.base" title="sklearn.base"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.base</span></code></a><a class="headerlink" href="#sklearn-base" title="Permalink to this headline">¶</a></h3>
<ul class="simple">
<li><p><span class="raw-html"><span class="badge badge-warning">API Change</span></span> From version 0.24 <a class="reference internal" href="../modules/generated/sklearn.base.BaseEstimator.html#sklearn.base.BaseEstimator.get_params" title="sklearn.base.BaseEstimator.get_params"><code class="xref py py-meth docutils literal notranslate"><span class="pre">base.BaseEstimator.get_params</span></code></a> will raise an
AttributeError rather than return None for parameters that are in the
estimator’s constructor but not stored as attributes on the instance.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/14464">#14464</a> by <a class="reference external" href="https://joelnothman.com/">Joel Nothman</a>.</p></li>
</ul>
</div>
<div class="section" id="sklearn-calibration">
<h3><a class="reference internal" href="../modules/classes.html#module-sklearn.calibration" title="sklearn.calibration"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.calibration</span></code></a><a class="headerlink" href="#sklearn-calibration" title="Permalink to this headline">¶</a></h3>
<ul class="simple">
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> Fixed a bug that made <a class="reference internal" href="../modules/generated/sklearn.calibration.CalibratedClassifierCV.html#sklearn.calibration.CalibratedClassifierCV" title="sklearn.calibration.CalibratedClassifierCV"><code class="xref py py-class docutils literal notranslate"><span class="pre">calibration.CalibratedClassifierCV</span></code></a> fail when
given a <code class="docutils literal notranslate"><span class="pre">sample_weight</span></code> parameter of type <code class="docutils literal notranslate"><span class="pre">list</span></code> (in the case where
<code class="docutils literal notranslate"><span class="pre">sample_weights</span></code> are not supported by the wrapped estimator). <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13575">#13575</a>
by <a class="reference external" href="https://github.com/wdevazelhes">William de Vazelhes</a>.</p></li>
</ul>
</div>
<div class="section" id="id5">
<h3><a class="reference internal" href="../modules/classes.html#module-sklearn.cluster" title="sklearn.cluster"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.cluster</span></code></a><a class="headerlink" href="#id5" title="Permalink to this headline">¶</a></h3>
<ul class="simple">
<li><p><span class="raw-html"><span class="badge badge-success">Feature</span></span> <a class="reference internal" href="../modules/generated/sklearn.cluster.SpectralClustering.html#sklearn.cluster.SpectralClustering" title="sklearn.cluster.SpectralClustering"><code class="xref py py-class docutils literal notranslate"><span class="pre">cluster.SpectralClustering</span></code></a> now accepts precomputed sparse
neighbors graph as input. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/issues/10482">#10482</a> by <a class="reference external" href="https://github.com/TomDLT">Tom Dupre la Tour</a> and
<a class="reference external" href="https://github.com/thechargedneutron">Kumar Ashutosh</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-info">Enhancement</span></span> <a class="reference internal" href="../modules/generated/sklearn.cluster.SpectralClustering.html#sklearn.cluster.SpectralClustering" title="sklearn.cluster.SpectralClustering"><code class="xref py py-class docutils literal notranslate"><span class="pre">cluster.SpectralClustering</span></code></a> now accepts a <code class="docutils literal notranslate"><span class="pre">n_components</span></code>
parameter. This parameter extends <code class="docutils literal notranslate"><span class="pre">SpectralClustering</span></code> class functionality to
match <a class="reference internal" href="../modules/generated/sklearn.cluster.spectral_clustering.html#sklearn.cluster.spectral_clustering" title="sklearn.cluster.spectral_clustering"><code class="xref py py-meth docutils literal notranslate"><span class="pre">cluster.spectral_clustering</span></code></a>.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13726">#13726</a> by <a class="reference external" href="https://github.com/fdas3213">Shuzhe Xiao</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> Fixed a bug where <a class="reference internal" href="../modules/generated/sklearn.cluster.KMeans.html#sklearn.cluster.KMeans" title="sklearn.cluster.KMeans"><code class="xref py py-class docutils literal notranslate"><span class="pre">cluster.KMeans</span></code></a> produced inconsistent results
between <code class="docutils literal notranslate"><span class="pre">n_jobs=1</span></code> and <code class="docutils literal notranslate"><span class="pre">n_jobs>1</span></code> due to the handling of the random state.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/9288">#9288</a> by <a class="reference external" href="https://github.com/bryanyang0528">Bryan Yang</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> Fixed a bug where <code class="docutils literal notranslate"><span class="pre">elkan</span></code> algorithm in <a class="reference internal" href="../modules/generated/sklearn.cluster.KMeans.html#sklearn.cluster.KMeans" title="sklearn.cluster.KMeans"><code class="xref py py-class docutils literal notranslate"><span class="pre">cluster.KMeans</span></code></a> was
producing Segmentation Fault on large arrays due to integer index overflow.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/15057">#15057</a> by <a class="reference external" href="https://github.com/balodja">Vladimir Korolev</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> <a class="reference internal" href="../modules/generated/sklearn.cluster.MeanShift.html#sklearn.cluster.MeanShift" title="sklearn.cluster.MeanShift"><code class="xref py py-class docutils literal notranslate"><span class="pre">MeanShift</span></code></a> now accepts a <a class="reference internal" href="../glossary.html#term-max-iter"><span class="xref std std-term">max_iter</span></a> with a
default value of 300 instead of always using the default 300. It also now
exposes an <code class="docutils literal notranslate"><span class="pre">n_iter_</span></code> indicating the maximum number of iterations performed
on each seed. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/15120">#15120</a> by <a class="reference external" href="https://github.com/adrinjalali">Adrin Jalali</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> <a class="reference internal" href="../modules/generated/sklearn.cluster.AgglomerativeClustering.html#sklearn.cluster.AgglomerativeClustering" title="sklearn.cluster.AgglomerativeClustering"><code class="xref py py-class docutils literal notranslate"><span class="pre">cluster.AgglomerativeClustering</span></code></a> and
<a class="reference internal" href="../modules/generated/sklearn.cluster.FeatureAgglomeration.html#sklearn.cluster.FeatureAgglomeration" title="sklearn.cluster.FeatureAgglomeration"><code class="xref py py-class docutils literal notranslate"><span class="pre">cluster.FeatureAgglomeration</span></code></a> now raise an error if
<code class="docutils literal notranslate"><span class="pre">affinity='cosine'</span></code> and <code class="docutils literal notranslate"><span class="pre">X</span></code> has samples that are all-zeros. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/7943">#7943</a> by
<a class="reference external" href="https://github.com/mthorrell">@mthorrell</a>.</p></li>
</ul>
</div>
<div class="section" id="sklearn-compose">
<h3><a class="reference internal" href="../modules/classes.html#module-sklearn.compose" title="sklearn.compose"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.compose</span></code></a><a class="headerlink" href="#sklearn-compose" title="Permalink to this headline">¶</a></h3>
<ul class="simple">
<li><p><span class="raw-html"><span class="badge badge-success">Feature</span></span> Adds <a class="reference internal" href="../modules/generated/sklearn.compose.make_column_selector.html#sklearn.compose.make_column_selector" title="sklearn.compose.make_column_selector"><code class="xref py py-func docutils literal notranslate"><span class="pre">compose.make_column_selector</span></code></a> which is used with
<a class="reference internal" href="../modules/generated/sklearn.compose.ColumnTransformer.html#sklearn.compose.ColumnTransformer" title="sklearn.compose.ColumnTransformer"><code class="xref py py-class docutils literal notranslate"><span class="pre">compose.ColumnTransformer</span></code></a> to select DataFrame columns on the basis
of name and dtype. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/12303">#12303</a> by <a class="reference external" href="https://github.com/thomasjpfan">Thomas Fan</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> Fixed a bug in <a class="reference internal" href="../modules/generated/sklearn.compose.ColumnTransformer.html#sklearn.compose.ColumnTransformer" title="sklearn.compose.ColumnTransformer"><code class="xref py py-class docutils literal notranslate"><span class="pre">compose.ColumnTransformer</span></code></a> which failed to
select the proper columns when using a boolean list, with NumPy older than
1.12.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/14510">#14510</a> by <a class="reference external" href="https://github.com/glemaitre">Guillaume Lemaitre</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> Fixed a bug in <a class="reference internal" href="../modules/generated/sklearn.compose.TransformedTargetRegressor.html#sklearn.compose.TransformedTargetRegressor" title="sklearn.compose.TransformedTargetRegressor"><code class="xref py py-class docutils literal notranslate"><span class="pre">compose.TransformedTargetRegressor</span></code></a> which did not
pass <code class="docutils literal notranslate"><span class="pre">**fit_params</span></code> to the underlying regressor.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/14890">#14890</a> by <a class="reference external" href="https://github.com/mfcabrera">Miguel Cabrera</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> The <a class="reference internal" href="../modules/generated/sklearn.compose.ColumnTransformer.html#sklearn.compose.ColumnTransformer" title="sklearn.compose.ColumnTransformer"><code class="xref py py-class docutils literal notranslate"><span class="pre">compose.ColumnTransformer</span></code></a> now requires the number of
features to be consistent between <code class="docutils literal notranslate"><span class="pre">fit</span></code> and <code class="docutils literal notranslate"><span class="pre">transform</span></code>. A <code class="docutils literal notranslate"><span class="pre">FutureWarning</span></code>
is raised now, and this will raise an error in 0.24. If the number of
features isn’t consistent and negative indexing is used, an error is
raised. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/14544">#14544</a> by <a class="reference external" href="https://github.com/adrinjalali">Adrin Jalali</a>.</p></li>
</ul>
</div>
<div class="section" id="sklearn-cross-decomposition">
<h3><a class="reference internal" href="../modules/classes.html#module-sklearn.cross_decomposition" title="sklearn.cross_decomposition"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.cross_decomposition</span></code></a><a class="headerlink" href="#sklearn-cross-decomposition" title="Permalink to this headline">¶</a></h3>
<ul class="simple">
<li><p><span class="raw-html"><span class="badge badge-success">Feature</span></span> <a class="reference internal" href="../modules/generated/sklearn.cross_decomposition.PLSCanonical.html#sklearn.cross_decomposition.PLSCanonical" title="sklearn.cross_decomposition.PLSCanonical"><code class="xref py py-class docutils literal notranslate"><span class="pre">cross_decomposition.PLSCanonical</span></code></a> and
<a class="reference internal" href="../modules/generated/sklearn.cross_decomposition.PLSRegression.html#sklearn.cross_decomposition.PLSRegression" title="sklearn.cross_decomposition.PLSRegression"><code class="xref py py-class docutils literal notranslate"><span class="pre">cross_decomposition.PLSRegression</span></code></a> have a new function
<code class="docutils literal notranslate"><span class="pre">inverse_transform</span></code> to transform data to the original space.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/15304">#15304</a> by <a class="reference external" href="https://github.com/jiwidi">Jaime Ferrando Huertas</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-info">Enhancement</span></span> <a class="reference internal" href="../modules/generated/sklearn.decomposition.KernelPCA.html#sklearn.decomposition.KernelPCA" title="sklearn.decomposition.KernelPCA"><code class="xref py py-class docutils literal notranslate"><span class="pre">decomposition.KernelPCA</span></code></a> now properly checks the
eigenvalues found by the solver for numerical or conditioning issues. This
ensures consistency of results across solvers (different choices for
<code class="docutils literal notranslate"><span class="pre">eigen_solver</span></code>), including approximate solvers such as <code class="docutils literal notranslate"><span class="pre">'randomized'</span></code> and
<code class="docutils literal notranslate"><span class="pre">'lobpcg'</span></code> (see <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/issues/12068">#12068</a>).
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/12145">#12145</a> by <a class="reference external" href="https://github.com/smarie">Sylvain Marié</a></p></li>
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> Fixed a bug where <a class="reference internal" href="../modules/generated/sklearn.cross_decomposition.PLSCanonical.html#sklearn.cross_decomposition.PLSCanonical" title="sklearn.cross_decomposition.PLSCanonical"><code class="xref py py-class docutils literal notranslate"><span class="pre">cross_decomposition.PLSCanonical</span></code></a> and
<a class="reference internal" href="../modules/generated/sklearn.cross_decomposition.PLSRegression.html#sklearn.cross_decomposition.PLSRegression" title="sklearn.cross_decomposition.PLSRegression"><code class="xref py py-class docutils literal notranslate"><span class="pre">cross_decomposition.PLSRegression</span></code></a> were raising an error when fitted
with a target matrix <code class="docutils literal notranslate"><span class="pre">Y</span></code> in which the first column was constant.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/issues/13609">#13609</a> by <a class="reference external" href="https://github.com/camilaagw">Camila Williamson</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> <a class="reference internal" href="../modules/generated/sklearn.cross_decomposition.CCA.html#sklearn.cross_decomposition.CCA" title="sklearn.cross_decomposition.CCA"><code class="xref py py-class docutils literal notranslate"><span class="pre">cross_decomposition.CCA</span></code></a> now produces the same results with
scipy 1.3 and previous scipy versions. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/15661">#15661</a> by <a class="reference external" href="https://github.com/thomasjpfan">Thomas Fan</a>.</p></li>
</ul>
</div>
<div class="section" id="sklearn-datasets">
<h3><a class="reference internal" href="../modules/classes.html#module-sklearn.datasets" title="sklearn.datasets"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.datasets</span></code></a><a class="headerlink" href="#sklearn-datasets" title="Permalink to this headline">¶</a></h3>
<ul class="simple">
<li><p><span class="raw-html"><span class="badge badge-success">Feature</span></span> <a class="reference internal" href="../modules/generated/sklearn.datasets.fetch_openml.html#sklearn.datasets.fetch_openml" title="sklearn.datasets.fetch_openml"><code class="xref py py-func docutils literal notranslate"><span class="pre">datasets.fetch_openml</span></code></a> now supports heterogeneous data using
pandas by setting <code class="docutils literal notranslate"><span class="pre">as_frame=True</span></code>. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13902">#13902</a> by <a class="reference external" href="https://github.com/thomasjpfan">Thomas Fan</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-success">Feature</span></span> <a class="reference internal" href="../modules/generated/sklearn.datasets.fetch_openml.html#sklearn.datasets.fetch_openml" title="sklearn.datasets.fetch_openml"><code class="xref py py-func docutils literal notranslate"><span class="pre">datasets.fetch_openml</span></code></a> now includes the <code class="docutils literal notranslate"><span class="pre">target_names</span></code> in
the returned Bunch. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/15160">#15160</a> by <a class="reference external" href="https://github.com/thomasjpfan">Thomas Fan</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-info">Enhancement</span></span> The parameter <code class="docutils literal notranslate"><span class="pre">return_X_y</span></code> was added to
<a class="reference internal" href="../modules/generated/sklearn.datasets.fetch_20newsgroups.html#sklearn.datasets.fetch_20newsgroups" title="sklearn.datasets.fetch_20newsgroups"><code class="xref py py-func docutils literal notranslate"><span class="pre">datasets.fetch_20newsgroups</span></code></a> and <a class="reference internal" href="../modules/generated/sklearn.datasets.fetch_olivetti_faces.html#sklearn.datasets.fetch_olivetti_faces" title="sklearn.datasets.fetch_olivetti_faces"><code class="xref py py-func docutils literal notranslate"><span class="pre">datasets.fetch_olivetti_faces</span></code></a>
. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/14259">#14259</a> by <a class="reference external" href="https://github.com/souravsingh">Sourav Singh</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-info">Enhancement</span></span> <a class="reference internal" href="../modules/generated/sklearn.datasets.make_classification.html#sklearn.datasets.make_classification" title="sklearn.datasets.make_classification"><code class="xref py py-func docutils literal notranslate"><span class="pre">datasets.make_classification</span></code></a> now accepts array-like
<code class="docutils literal notranslate"><span class="pre">weights</span></code> parameter, i.e. list or numpy.array, instead of list only.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/14764">#14764</a> by <a class="reference external" href="https://github.com/CatChenal">Cat Chenal</a>.</p></li>
<li><dl class="simple">
<dt><span class="raw-html"><span class="badge badge-info">Enhancement</span></span> The parameter <code class="docutils literal notranslate"><span class="pre">normalize</span></code> was added to</dt><dd><p><a class="reference internal" href="../modules/generated/sklearn.datasets.fetch_20newsgroups_vectorized.html#sklearn.datasets.fetch_20newsgroups_vectorized" title="sklearn.datasets.fetch_20newsgroups_vectorized"><code class="xref py py-func docutils literal notranslate"><span class="pre">datasets.fetch_20newsgroups_vectorized</span></code></a>.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/14740">#14740</a> by <a class="reference external" href="https://github.com/stephantul">Stéphan Tulkens</a></p>
</dd>
</dl>
</li>
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> Fixed a bug in <a class="reference internal" href="../modules/generated/sklearn.datasets.fetch_openml.html#sklearn.datasets.fetch_openml" title="sklearn.datasets.fetch_openml"><code class="xref py py-func docutils literal notranslate"><span class="pre">datasets.fetch_openml</span></code></a>, which failed to load
an OpenML dataset that contains an ignored feature.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/14623">#14623</a> by <a class="reference external" href="https://github.com/HabchiSarra">Sarra Habchi</a>.</p></li>
</ul>
</div>
<div class="section" id="sklearn-decomposition">
<h3><a class="reference internal" href="../modules/classes.html#module-sklearn.decomposition" title="sklearn.decomposition"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.decomposition</span></code></a><a class="headerlink" href="#sklearn-decomposition" title="Permalink to this headline">¶</a></h3>
<ul class="simple">
<li><p><span class="raw-html"><span class="badge badge-info">Efficiency</span></span> <code class="xref py py-class docutils literal notranslate"><span class="pre">decomposition.NMF(solver='mu')</span></code> fitted on sparse input
matrices now uses batching to avoid briefly allocating an array with size
(#non-zero elements, n_components). <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/15257">#15257</a> by <a class="reference external" href="Maocx">Mart Willocx</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-info">Enhancement</span></span> <a class="reference internal" href="../modules/generated/sklearn.decomposition.dict_learning.html#sklearn.decomposition.dict_learning" title="sklearn.decomposition.dict_learning"><code class="xref py py-func docutils literal notranslate"><span class="pre">decomposition.dict_learning</span></code></a> and
<a class="reference internal" href="../modules/generated/sklearn.decomposition.dict_learning_online.html#sklearn.decomposition.dict_learning_online" title="sklearn.decomposition.dict_learning_online"><code class="xref py py-func docutils literal notranslate"><span class="pre">decomposition.dict_learning_online</span></code></a> now accept <code class="docutils literal notranslate"><span class="pre">method_max_iter</span></code> and
pass it to <a class="reference internal" href="../modules/generated/sklearn.decomposition.sparse_encode.html#sklearn.decomposition.sparse_encode" title="sklearn.decomposition.sparse_encode"><code class="xref py py-meth docutils literal notranslate"><span class="pre">decomposition.sparse_encode</span></code></a>.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/issues/12650">#12650</a> by <a class="reference external" href="https://github.com/adrinjalali">Adrin Jalali</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-info">Enhancement</span></span> <a class="reference internal" href="../modules/generated/sklearn.decomposition.SparseCoder.html#sklearn.decomposition.SparseCoder" title="sklearn.decomposition.SparseCoder"><code class="xref py py-class docutils literal notranslate"><span class="pre">decomposition.SparseCoder</span></code></a>,
<a class="reference internal" href="../modules/generated/sklearn.decomposition.DictionaryLearning.html#sklearn.decomposition.DictionaryLearning" title="sklearn.decomposition.DictionaryLearning"><code class="xref py py-class docutils literal notranslate"><span class="pre">decomposition.DictionaryLearning</span></code></a>, and
<a class="reference internal" href="../modules/generated/sklearn.decomposition.MiniBatchDictionaryLearning.html#sklearn.decomposition.MiniBatchDictionaryLearning" title="sklearn.decomposition.MiniBatchDictionaryLearning"><code class="xref py py-class docutils literal notranslate"><span class="pre">decomposition.MiniBatchDictionaryLearning</span></code></a> now take a
<code class="docutils literal notranslate"><span class="pre">transform_max_iter</span></code> parameter and pass it to either
<a class="reference internal" href="../modules/generated/sklearn.decomposition.dict_learning.html#sklearn.decomposition.dict_learning" title="sklearn.decomposition.dict_learning"><code class="xref py py-func docutils literal notranslate"><span class="pre">decomposition.dict_learning</span></code></a> or
<a class="reference internal" href="../modules/generated/sklearn.decomposition.sparse_encode.html#sklearn.decomposition.sparse_encode" title="sklearn.decomposition.sparse_encode"><code class="xref py py-func docutils literal notranslate"><span class="pre">decomposition.sparse_encode</span></code></a>. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/issues/12650">#12650</a> by <a class="reference external" href="https://github.com/adrinjalali">Adrin Jalali</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-info">Enhancement</span></span> <a class="reference internal" href="../modules/generated/sklearn.decomposition.IncrementalPCA.html#sklearn.decomposition.IncrementalPCA" title="sklearn.decomposition.IncrementalPCA"><code class="xref py py-class docutils literal notranslate"><span class="pre">decomposition.IncrementalPCA</span></code></a> now accepts sparse
matrices as input, converting them to dense in batches thereby avoiding the
need to store the entire dense matrix at once.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13960">#13960</a> by <a class="reference external" href="https://github.com/scottgigante">Scott Gigante</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> <a class="reference internal" href="../modules/generated/sklearn.decomposition.sparse_encode.html#sklearn.decomposition.sparse_encode" title="sklearn.decomposition.sparse_encode"><code class="xref py py-func docutils literal notranslate"><span class="pre">decomposition.sparse_encode</span></code></a> now passes the <code class="docutils literal notranslate"><span class="pre">max_iter</span></code> to the
underlying <a class="reference internal" href="../modules/generated/sklearn.linear_model.LassoLars.html#sklearn.linear_model.LassoLars" title="sklearn.linear_model.LassoLars"><code class="xref py py-class docutils literal notranslate"><span class="pre">linear_model.LassoLars</span></code></a> when <code class="docutils literal notranslate"><span class="pre">algorithm='lasso_lars'</span></code>.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/issues/12650">#12650</a> by <a class="reference external" href="https://github.com/adrinjalali">Adrin Jalali</a>.</p></li>
</ul>
</div>
<div class="section" id="sklearn-dummy">
<h3><a class="reference internal" href="../modules/classes.html#module-sklearn.dummy" title="sklearn.dummy"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.dummy</span></code></a><a class="headerlink" href="#sklearn-dummy" title="Permalink to this headline">¶</a></h3>
<ul class="simple">
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> <a class="reference internal" href="../modules/generated/sklearn.dummy.DummyClassifier.html#sklearn.dummy.DummyClassifier" title="sklearn.dummy.DummyClassifier"><code class="xref py py-class docutils literal notranslate"><span class="pre">dummy.DummyClassifier</span></code></a> now handles checking the existence
of the provided constant in multiouput cases.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/14908">#14908</a> by <a class="reference external" href="https://github.com/martinagvilas">Martina G. Vilas</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-warning">API Change</span></span> The default value of the <code class="docutils literal notranslate"><span class="pre">strategy</span></code> parameter in
<a class="reference internal" href="../modules/generated/sklearn.dummy.DummyClassifier.html#sklearn.dummy.DummyClassifier" title="sklearn.dummy.DummyClassifier"><code class="xref py py-class docutils literal notranslate"><span class="pre">dummy.DummyClassifier</span></code></a> will change from <code class="docutils literal notranslate"><span class="pre">'stratified'</span></code> in version
0.22 to <code class="docutils literal notranslate"><span class="pre">'prior'</span></code> in 0.24. A FutureWarning is raised when the default value
is used. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/15382">#15382</a> by <a class="reference external" href="https://github.com/thomasjpfan">Thomas Fan</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-warning">API Change</span></span> The <code class="docutils literal notranslate"><span class="pre">outputs_2d_</span></code> attribute is deprecated in
<a class="reference internal" href="../modules/generated/sklearn.dummy.DummyClassifier.html#sklearn.dummy.DummyClassifier" title="sklearn.dummy.DummyClassifier"><code class="xref py py-class docutils literal notranslate"><span class="pre">dummy.DummyClassifier</span></code></a> and <a class="reference internal" href="../modules/generated/sklearn.dummy.DummyRegressor.html#sklearn.dummy.DummyRegressor" title="sklearn.dummy.DummyRegressor"><code class="xref py py-class docutils literal notranslate"><span class="pre">dummy.DummyRegressor</span></code></a>. It is
equivalent to <code class="docutils literal notranslate"><span class="pre">n_outputs</span> <span class="pre">></span> <span class="pre">1</span></code>. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/14933">#14933</a> by <a class="reference external" href="https://github.com/NicolasHug">Nicolas Hug</a></p></li>
</ul>
</div>
<div class="section" id="sklearn-ensemble">
<h3><a class="reference internal" href="../modules/classes.html#module-sklearn.ensemble" title="sklearn.ensemble"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.ensemble</span></code></a><a class="headerlink" href="#sklearn-ensemble" title="Permalink to this headline">¶</a></h3>
<ul>
<li><p><span class="raw-html"><span class="badge badge-success">Major Feature</span></span> Added <a class="reference internal" href="../modules/generated/sklearn.ensemble.StackingClassifier.html#sklearn.ensemble.StackingClassifier" title="sklearn.ensemble.StackingClassifier"><code class="xref py py-class docutils literal notranslate"><span class="pre">ensemble.StackingClassifier</span></code></a> and
<a class="reference internal" href="../modules/generated/sklearn.ensemble.StackingRegressor.html#sklearn.ensemble.StackingRegressor" title="sklearn.ensemble.StackingRegressor"><code class="xref py py-class docutils literal notranslate"><span class="pre">ensemble.StackingRegressor</span></code></a> to stack predictors using a final
classifier or regressor. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/11047">#11047</a> by <a class="reference external" href="https://github.com/glemaitre">Guillaume Lemaitre</a> and <a class="reference external" href="https://github.com/caioaao">Caio Oliveira</a> and <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/15138">#15138</a> by
<a class="reference external" href="https://github.com/jcusick13">Jon Cusick</a>..</p></li>
<li><p>Many improvements were made to
<a class="reference internal" href="../modules/generated/sklearn.ensemble.HistGradientBoostingClassifier.html#sklearn.ensemble.HistGradientBoostingClassifier" title="sklearn.ensemble.HistGradientBoostingClassifier"><code class="xref py py-class docutils literal notranslate"><span class="pre">ensemble.HistGradientBoostingClassifier</span></code></a> and
<a class="reference internal" href="../modules/generated/sklearn.ensemble.HistGradientBoostingRegressor.html#sklearn.ensemble.HistGradientBoostingRegressor" title="sklearn.ensemble.HistGradientBoostingRegressor"><code class="xref py py-class docutils literal notranslate"><span class="pre">ensemble.HistGradientBoostingRegressor</span></code></a>:</p>
<ul class="simple">
<li><p><span class="raw-html"><span class="badge badge-success">Major Feature</span></span> Estimators now natively support dense data with missing
values both for training and predicting. They also support infinite
values. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13911">#13911</a> and <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/14406">#14406</a> by <a class="reference external" href="https://github.com/NicolasHug">Nicolas Hug</a>, <a class="reference external" href="https://github.com/adrinjalali">Adrin Jalali</a>
and <a class="reference external" href="https://twitter.com/ogrisel">Olivier Grisel</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-success">Feature</span></span> Estimators now have an additional <code class="docutils literal notranslate"><span class="pre">warm_start</span></code> parameter that
enables warm starting. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/14012">#14012</a> by <a class="reference external" href="https://github.com/johannfaouzi">Johann Faouzi</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-success">Feature</span></span> <a class="reference internal" href="../modules/generated/sklearn.inspection.partial_dependence.html#sklearn.inspection.partial_dependence" title="sklearn.inspection.partial_dependence"><code class="xref py py-func docutils literal notranslate"><span class="pre">inspection.partial_dependence</span></code></a> and
<a class="reference internal" href="../modules/generated/sklearn.inspection.plot_partial_dependence.html#sklearn.inspection.plot_partial_dependence" title="sklearn.inspection.plot_partial_dependence"><code class="xref py py-func docutils literal notranslate"><span class="pre">inspection.plot_partial_dependence</span></code></a> now support the fast ‘recursion’
method for both estimators. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13769">#13769</a> by <a class="reference external" href="https://github.com/NicolasHug">Nicolas Hug</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-info">Enhancement</span></span> for <a class="reference internal" href="../modules/generated/sklearn.ensemble.HistGradientBoostingClassifier.html#sklearn.ensemble.HistGradientBoostingClassifier" title="sklearn.ensemble.HistGradientBoostingClassifier"><code class="xref py py-class docutils literal notranslate"><span class="pre">ensemble.HistGradientBoostingClassifier</span></code></a> the
training loss or score is now monitored on a class-wise stratified
subsample to preserve the class balance of the original training set.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/14194">#14194</a> by <a class="reference external" href="https://github.com/johannfaouzi">Johann Faouzi</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-info">Enhancement</span></span> <a class="reference internal" href="../modules/generated/sklearn.ensemble.HistGradientBoostingRegressor.html#sklearn.ensemble.HistGradientBoostingRegressor" title="sklearn.ensemble.HistGradientBoostingRegressor"><code class="xref py py-class docutils literal notranslate"><span class="pre">ensemble.HistGradientBoostingRegressor</span></code></a> now supports
the ‘least_absolute_deviation’ loss. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13896">#13896</a> by <a class="reference external" href="https://github.com/NicolasHug">Nicolas Hug</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> Estimators now bin the training and validation data separately to
avoid any data leak. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13933">#13933</a> by <a class="reference external" href="https://github.com/NicolasHug">Nicolas Hug</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> Fixed a bug where early stopping would break with string targets.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/14710">#14710</a> by <a class="reference external" href="https://github.com/glemaitre">Guillaume Lemaitre</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> <a class="reference internal" href="../modules/generated/sklearn.ensemble.HistGradientBoostingClassifier.html#sklearn.ensemble.HistGradientBoostingClassifier" title="sklearn.ensemble.HistGradientBoostingClassifier"><code class="xref py py-class docutils literal notranslate"><span class="pre">ensemble.HistGradientBoostingClassifier</span></code></a> now raises an error
if <code class="docutils literal notranslate"><span class="pre">categorical_crossentropy</span></code> loss is given for a binary classification
problem. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/14869">#14869</a> by <a class="reference external" href="https://github.com/adrinjalali">Adrin Jalali</a>.</p></li>
</ul>
<p>Note that pickles from 0.21 will not work in 0.22.</p>
</li>
<li><p><span class="raw-html"><span class="badge badge-info">Enhancement</span></span> Addition of <code class="docutils literal notranslate"><span class="pre">max_samples</span></code> argument allows limiting
size of bootstrap samples to be less than size of dataset. Added to
<a class="reference internal" href="../modules/generated/sklearn.ensemble.RandomForestClassifier.html#sklearn.ensemble.RandomForestClassifier" title="sklearn.ensemble.RandomForestClassifier"><code class="xref py py-class docutils literal notranslate"><span class="pre">ensemble.RandomForestClassifier</span></code></a>,
<a class="reference internal" href="../modules/generated/sklearn.ensemble.RandomForestRegressor.html#sklearn.ensemble.RandomForestRegressor" title="sklearn.ensemble.RandomForestRegressor"><code class="xref py py-class docutils literal notranslate"><span class="pre">ensemble.RandomForestRegressor</span></code></a>,
<a class="reference internal" href="../modules/generated/sklearn.ensemble.ExtraTreesClassifier.html#sklearn.ensemble.ExtraTreesClassifier" title="sklearn.ensemble.ExtraTreesClassifier"><code class="xref py py-class docutils literal notranslate"><span class="pre">ensemble.ExtraTreesClassifier</span></code></a>,
<a class="reference internal" href="../modules/generated/sklearn.ensemble.ExtraTreesRegressor.html#sklearn.ensemble.ExtraTreesRegressor" title="sklearn.ensemble.ExtraTreesRegressor"><code class="xref py py-class docutils literal notranslate"><span class="pre">ensemble.ExtraTreesRegressor</span></code></a>. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/14682">#14682</a> by
<a class="reference external" href="https://github.com/notmatthancock">Matt Hancock</a> and
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/5963">#5963</a> by <a class="reference external" href="https://github.com/DrDub">Pablo Duboue</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> <a class="reference internal" href="../modules/generated/sklearn.ensemble.VotingClassifier.html#sklearn.ensemble.VotingClassifier.predict_proba" title="sklearn.ensemble.VotingClassifier.predict_proba"><code class="xref py py-func docutils literal notranslate"><span class="pre">ensemble.VotingClassifier.predict_proba</span></code></a> will no longer be
present when <code class="docutils literal notranslate"><span class="pre">voting='hard'</span></code>. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/14287">#14287</a> by <a class="reference external" href="https://github.com/thomasjpfan">Thomas Fan</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> The <code class="docutils literal notranslate"><span class="pre">named_estimators_</span></code> attribute in <a class="reference internal" href="../modules/generated/sklearn.ensemble.VotingClassifier.html#sklearn.ensemble.VotingClassifier" title="sklearn.ensemble.VotingClassifier"><code class="xref py py-class docutils literal notranslate"><span class="pre">ensemble.VotingClassifier</span></code></a>
and <a class="reference internal" href="../modules/generated/sklearn.ensemble.VotingRegressor.html#sklearn.ensemble.VotingRegressor" title="sklearn.ensemble.VotingRegressor"><code class="xref py py-class docutils literal notranslate"><span class="pre">ensemble.VotingRegressor</span></code></a> now correctly maps to dropped estimators.
Previously, the <code class="docutils literal notranslate"><span class="pre">named_estimators_</span></code> mapping was incorrect whenever one of the
estimators was dropped. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/15375">#15375</a> by <a class="reference external" href="https://github.com/thomasjpfan">Thomas Fan</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> Run by default
<a class="reference internal" href="../modules/generated/sklearn.utils.estimator_checks.check_estimator.html#sklearn.utils.estimator_checks.check_estimator" title="sklearn.utils.estimator_checks.check_estimator"><code class="xref py py-func docutils literal notranslate"><span class="pre">utils.estimator_checks.check_estimator</span></code></a> on both
<a class="reference internal" href="../modules/generated/sklearn.ensemble.VotingClassifier.html#sklearn.ensemble.VotingClassifier" title="sklearn.ensemble.VotingClassifier"><code class="xref py py-class docutils literal notranslate"><span class="pre">ensemble.VotingClassifier</span></code></a> and <a class="reference internal" href="../modules/generated/sklearn.ensemble.VotingRegressor.html#sklearn.ensemble.VotingRegressor" title="sklearn.ensemble.VotingRegressor"><code class="xref py py-class docutils literal notranslate"><span class="pre">ensemble.VotingRegressor</span></code></a>. It
leads to solve issues regarding shape consistency during <code class="docutils literal notranslate"><span class="pre">predict</span></code> which was
failing when the underlying estimators were not outputting consistent array
dimensions. Note that it should be replaced by refactoring the common tests
in the future.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/14305">#14305</a> by <a class="reference external" href="https://github.com/glemaitre">Guillaume Lemaitre</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> <a class="reference internal" href="../modules/generated/sklearn.ensemble.AdaBoostClassifier.html#sklearn.ensemble.AdaBoostClassifier" title="sklearn.ensemble.AdaBoostClassifier"><code class="xref py py-class docutils literal notranslate"><span class="pre">ensemble.AdaBoostClassifier</span></code></a> computes probabilities based on
the decision function as in the literature. Thus, <code class="docutils literal notranslate"><span class="pre">predict</span></code> and
<code class="docutils literal notranslate"><span class="pre">predict_proba</span></code> give consistent results.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/14114">#14114</a> by <a class="reference external" href="https://github.com/glemaitre">Guillaume Lemaitre</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> Stacking and Voting estimators now ensure that their underlying
estimators are either all classifiers or all regressors.
<a class="reference internal" href="../modules/generated/sklearn.ensemble.StackingClassifier.html#sklearn.ensemble.StackingClassifier" title="sklearn.ensemble.StackingClassifier"><code class="xref py py-class docutils literal notranslate"><span class="pre">ensemble.StackingClassifier</span></code></a>, <a class="reference internal" href="../modules/generated/sklearn.ensemble.StackingRegressor.html#sklearn.ensemble.StackingRegressor" title="sklearn.ensemble.StackingRegressor"><code class="xref py py-class docutils literal notranslate"><span class="pre">ensemble.StackingRegressor</span></code></a>,
and <a class="reference internal" href="../modules/generated/sklearn.ensemble.VotingClassifier.html#sklearn.ensemble.VotingClassifier" title="sklearn.ensemble.VotingClassifier"><code class="xref py py-class docutils literal notranslate"><span class="pre">ensemble.VotingClassifier</span></code></a> and <code class="xref py py-class docutils literal notranslate"><span class="pre">VotingRegressor</span></code>
now raise consistent error messages.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/15084">#15084</a> by <a class="reference external" href="https://github.com/glemaitre">Guillaume Lemaitre</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> <a class="reference internal" href="../modules/generated/sklearn.ensemble.AdaBoostRegressor.html#sklearn.ensemble.AdaBoostRegressor" title="sklearn.ensemble.AdaBoostRegressor"><code class="xref py py-class docutils literal notranslate"><span class="pre">ensemble.AdaBoostRegressor</span></code></a> where the loss should be normalized
by the max of the samples with non-null weights only.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/14294">#14294</a> by <a class="reference external" href="https://github.com/glemaitre">Guillaume Lemaitre</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-warning">API Change</span></span> <code class="docutils literal notranslate"><span class="pre">presort</span></code> is now deprecated in
<a class="reference internal" href="../modules/generated/sklearn.ensemble.GradientBoostingClassifier.html#sklearn.ensemble.GradientBoostingClassifier" title="sklearn.ensemble.GradientBoostingClassifier"><code class="xref py py-class docutils literal notranslate"><span class="pre">ensemble.GradientBoostingClassifier</span></code></a> and
<a class="reference internal" href="../modules/generated/sklearn.ensemble.GradientBoostingRegressor.html#sklearn.ensemble.GradientBoostingRegressor" title="sklearn.ensemble.GradientBoostingRegressor"><code class="xref py py-class docutils literal notranslate"><span class="pre">ensemble.GradientBoostingRegressor</span></code></a>, and the parameter has no effect.
Users are recommended to use <a class="reference internal" href="../modules/generated/sklearn.ensemble.HistGradientBoostingClassifier.html#sklearn.ensemble.HistGradientBoostingClassifier" title="sklearn.ensemble.HistGradientBoostingClassifier"><code class="xref py py-class docutils literal notranslate"><span class="pre">ensemble.HistGradientBoostingClassifier</span></code></a>
and <a class="reference internal" href="../modules/generated/sklearn.ensemble.HistGradientBoostingRegressor.html#sklearn.ensemble.HistGradientBoostingRegressor" title="sklearn.ensemble.HistGradientBoostingRegressor"><code class="xref py py-class docutils literal notranslate"><span class="pre">ensemble.HistGradientBoostingRegressor</span></code></a> instead.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/14907">#14907</a> by <a class="reference external" href="https://github.com/adrinjalali">Adrin Jalali</a>.</p></li>
</ul>
</div>
<div class="section" id="sklearn-feature-extraction">
<h3><a class="reference internal" href="../modules/classes.html#module-sklearn.feature_extraction" title="sklearn.feature_extraction"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.feature_extraction</span></code></a><a class="headerlink" href="#sklearn-feature-extraction" title="Permalink to this headline">¶</a></h3>
<ul class="simple">
<li><p><span class="raw-html"><span class="badge badge-info">Enhancement</span></span> A warning will now be raised if a parameter choice means
that another parameter will be unused on calling the fit() method for
<a class="reference internal" href="../modules/generated/sklearn.feature_extraction.text.HashingVectorizer.html#sklearn.feature_extraction.text.HashingVectorizer" title="sklearn.feature_extraction.text.HashingVectorizer"><code class="xref py py-class docutils literal notranslate"><span class="pre">feature_extraction.text.HashingVectorizer</span></code></a>,
<a class="reference internal" href="../modules/generated/sklearn.feature_extraction.text.CountVectorizer.html#sklearn.feature_extraction.text.CountVectorizer" title="sklearn.feature_extraction.text.CountVectorizer"><code class="xref py py-class docutils literal notranslate"><span class="pre">feature_extraction.text.CountVectorizer</span></code></a> and
<a class="reference internal" href="../modules/generated/sklearn.feature_extraction.text.TfidfVectorizer.html#sklearn.feature_extraction.text.TfidfVectorizer" title="sklearn.feature_extraction.text.TfidfVectorizer"><code class="xref py py-class docutils literal notranslate"><span class="pre">feature_extraction.text.TfidfVectorizer</span></code></a>.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/14602">#14602</a> by <a class="reference external" href="https://github.com/getgaurav2">Gaurav Chawla</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> Functions created by <code class="docutils literal notranslate"><span class="pre">build_preprocessor</span></code> and <code class="docutils literal notranslate"><span class="pre">build_analyzer</span></code> of
<code class="xref py py-class docutils literal notranslate"><span class="pre">feature_extraction.text.VectorizerMixin</span></code> can now be pickled.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/14430">#14430</a> by <a class="reference external" href="https://github.com/deniederhut">Dillon Niederhut</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> <code class="xref py py-func docutils literal notranslate"><span class="pre">feature_extraction.text.strip_accents_unicode</span></code> now correctly
removes accents from strings that are in NFKD normalized form. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/15100">#15100</a> by
<a class="reference external" href="https://github.com/DGrady">Daniel Grady</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> Fixed a bug that caused <a class="reference internal" href="../modules/generated/sklearn.feature_extraction.DictVectorizer.html#sklearn.feature_extraction.DictVectorizer" title="sklearn.feature_extraction.DictVectorizer"><code class="xref py py-class docutils literal notranslate"><span class="pre">feature_extraction.DictVectorizer</span></code></a> to raise
an <code class="docutils literal notranslate"><span class="pre">OverflowError</span></code> during the <code class="docutils literal notranslate"><span class="pre">transform</span></code> operation when producing a <code class="docutils literal notranslate"><span class="pre">scipy.sparse</span></code>
matrix on large input data. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/15463">#15463</a> by <a class="reference external" href="https://github.com/norvan">Norvan Sahiner</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-warning">API Change</span></span> Deprecated unused <code class="docutils literal notranslate"><span class="pre">copy</span></code> param for
<a class="reference internal" href="../modules/generated/sklearn.feature_extraction.text.TfidfVectorizer.html#sklearn.feature_extraction.text.TfidfVectorizer.transform" title="sklearn.feature_extraction.text.TfidfVectorizer.transform"><code class="xref py py-meth docutils literal notranslate"><span class="pre">feature_extraction.text.TfidfVectorizer.transform</span></code></a> it will be
removed in v0.24. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/14520">#14520</a> by
<a class="reference external" href="https://github.com/guillemgsubies">Guillem G. Subies</a>.</p></li>
</ul>
</div>
<div class="section" id="sklearn-feature-selection">
<h3><a class="reference internal" href="../modules/classes.html#module-sklearn.feature_selection" title="sklearn.feature_selection"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.feature_selection</span></code></a><a class="headerlink" href="#sklearn-feature-selection" title="Permalink to this headline">¶</a></h3>
<ul class="simple">
<li><p><span class="raw-html"><span class="badge badge-info">Enhancement</span></span> Updated the following <code class="xref py py-mod docutils literal notranslate"><span class="pre">feature_selection</span></code> estimators to allow
NaN/Inf values in <code class="docutils literal notranslate"><span class="pre">transform</span></code> and <code class="docutils literal notranslate"><span class="pre">fit</span></code>:
<a class="reference internal" href="../modules/generated/sklearn.feature_selection.RFE.html#sklearn.feature_selection.RFE" title="sklearn.feature_selection.RFE"><code class="xref py py-class docutils literal notranslate"><span class="pre">feature_selection.RFE</span></code></a>, <a class="reference internal" href="../modules/generated/sklearn.feature_selection.RFECV.html#sklearn.feature_selection.RFECV" title="sklearn.feature_selection.RFECV"><code class="xref py py-class docutils literal notranslate"><span class="pre">feature_selection.RFECV</span></code></a>,
<a class="reference internal" href="../modules/generated/sklearn.feature_selection.SelectFromModel.html#sklearn.feature_selection.SelectFromModel" title="sklearn.feature_selection.SelectFromModel"><code class="xref py py-class docutils literal notranslate"><span class="pre">feature_selection.SelectFromModel</span></code></a>,
and <a class="reference internal" href="../modules/generated/sklearn.feature_selection.VarianceThreshold.html#sklearn.feature_selection.VarianceThreshold" title="sklearn.feature_selection.VarianceThreshold"><code class="xref py py-class docutils literal notranslate"><span class="pre">feature_selection.VarianceThreshold</span></code></a>. Note that if the underlying
estimator of the feature selector does not allow NaN/Inf then it will still
error, but the feature selectors themselves no longer enforce this
restriction unnecessarily. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/issues/11635">#11635</a> by <a class="reference external" href="https://github.com/adpeters">Alec Peters</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> Fixed a bug where <a class="reference internal" href="../modules/generated/sklearn.feature_selection.VarianceThreshold.html#sklearn.feature_selection.VarianceThreshold" title="sklearn.feature_selection.VarianceThreshold"><code class="xref py py-class docutils literal notranslate"><span class="pre">feature_selection.VarianceThreshold</span></code></a> with
<code class="docutils literal notranslate"><span class="pre">threshold=0</span></code> did not remove constant features due to numerical instability,
by using range rather than variance in this case.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13704">#13704</a> by <a class="reference external" href="https://github.com/rlms">Roddy MacSween</a>.</p></li>
</ul>
</div>
<div class="section" id="sklearn-gaussian-process">
<h3><a class="reference internal" href="../modules/classes.html#module-sklearn.gaussian_process" title="sklearn.gaussian_process"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.gaussian_process</span></code></a><a class="headerlink" href="#sklearn-gaussian-process" title="Permalink to this headline">¶</a></h3>
<ul class="simple">
<li><p><span class="raw-html"><span class="badge badge-success">Feature</span></span> Gaussian process models on structured data: <a class="reference internal" href="../modules/generated/sklearn.gaussian_process.GaussianProcessRegressor.html#sklearn.gaussian_process.GaussianProcessRegressor" title="sklearn.gaussian_process.GaussianProcessRegressor"><code class="xref py py-class docutils literal notranslate"><span class="pre">gaussian_process.GaussianProcessRegressor</span></code></a>
and <a class="reference internal" href="../modules/generated/sklearn.gaussian_process.GaussianProcessClassifier.html#sklearn.gaussian_process.GaussianProcessClassifier" title="sklearn.gaussian_process.GaussianProcessClassifier"><code class="xref py py-class docutils literal notranslate"><span class="pre">gaussian_process.GaussianProcessClassifier</span></code></a> can now accept a list
of generic objects (e.g. strings, trees, graphs, etc.) as the <code class="docutils literal notranslate"><span class="pre">X</span></code> argument
to their training/prediction methods.
A user-defined kernel should be provided for computing the kernel matrix among
the generic objects, and should inherit from <code class="xref py py-class docutils literal notranslate"><span class="pre">gaussian_process.kernels.GenericKernelMixin</span></code>
to notify the GPR/GPC model that it handles non-vectorial samples.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/15557">#15557</a> by <a class="reference external" href="https://github.com/yhtang">Yu-Hang Tang</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-info">Efficiency</span></span> <a class="reference internal" href="../modules/generated/sklearn.gaussian_process.GaussianProcessClassifier.html#sklearn.gaussian_process.GaussianProcessClassifier.log_marginal_likelihood" title="sklearn.gaussian_process.GaussianProcessClassifier.log_marginal_likelihood"><code class="xref py py-func docutils literal notranslate"><span class="pre">gaussian_process.GaussianProcessClassifier.log_marginal_likelihood</span></code></a>
and <a class="reference internal" href="../modules/generated/sklearn.gaussian_process.GaussianProcessRegressor.html#sklearn.gaussian_process.GaussianProcessRegressor.log_marginal_likelihood" title="sklearn.gaussian_process.GaussianProcessRegressor.log_marginal_likelihood"><code class="xref py py-func docutils literal notranslate"><span class="pre">gaussian_process.GaussianProcessRegressor.log_marginal_likelihood</span></code></a> now
accept a <code class="docutils literal notranslate"><span class="pre">clone_kernel=True</span></code> keyword argument. When set to <code class="docutils literal notranslate"><span class="pre">False</span></code>,
the kernel attribute is modified, but may result in a performance improvement.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/14378">#14378</a> by <a class="reference external" href="https://github.com/c-bata">Masashi Shibata</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-warning">API Change</span></span> From version 0.24 <a class="reference internal" href="../modules/generated/sklearn.gaussian_process.kernels.Kernel.html#sklearn.gaussian_process.kernels.Kernel.get_params" title="sklearn.gaussian_process.kernels.Kernel.get_params"><code class="xref py py-meth docutils literal notranslate"><span class="pre">gaussian_process.kernels.Kernel.get_params</span></code></a> will raise an
<code class="docutils literal notranslate"><span class="pre">AttributeError</span></code> rather than return <code class="docutils literal notranslate"><span class="pre">None</span></code> for parameters that are in the
estimator’s constructor but not stored as attributes on the instance.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/14464">#14464</a> by <a class="reference external" href="https://joelnothman.com/">Joel Nothman</a>.</p></li>
</ul>
</div>
<div class="section" id="id6">
<h3><a class="reference internal" href="../modules/classes.html#module-sklearn.impute" title="sklearn.impute"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.impute</span></code></a><a class="headerlink" href="#id6" title="Permalink to this headline">¶</a></h3>
<ul class="simple">
<li><p><span class="raw-html"><span class="badge badge-success">Major Feature</span></span> Added <a class="reference internal" href="../modules/generated/sklearn.impute.KNNImputer.html#sklearn.impute.KNNImputer" title="sklearn.impute.KNNImputer"><code class="xref py py-class docutils literal notranslate"><span class="pre">impute.KNNImputer</span></code></a>, to impute missing values using
k-Nearest Neighbors. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/issues/12852">#12852</a> by <a class="reference external" href="https://github.com/ashimb9">Ashim Bhattarai</a> and
<a class="reference external" href="https://github.com/thomasjpfan">Thomas Fan</a> and <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/15010">#15010</a> by <a class="reference external" href="https://github.com/glemaitre">Guillaume Lemaitre</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-success">Feature</span></span> <a class="reference internal" href="../modules/generated/sklearn.impute.IterativeImputer.html#sklearn.impute.IterativeImputer" title="sklearn.impute.IterativeImputer"><code class="xref py py-class docutils literal notranslate"><span class="pre">impute.IterativeImputer</span></code></a> has new <code class="docutils literal notranslate"><span class="pre">skip_compute</span></code> flag that
is False by default, which, when True, will skip computation on features that
have no missing values during the fit phase. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/issues/13773">#13773</a> by
<a class="reference external" href="https://github.com/sergeyf">Sergey Feldman</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-info">Efficiency</span></span> <a class="reference internal" href="../modules/generated/sklearn.impute.MissingIndicator.html#sklearn.impute.MissingIndicator.fit_transform" title="sklearn.impute.MissingIndicator.fit_transform"><code class="xref py py-meth docutils literal notranslate"><span class="pre">impute.MissingIndicator.fit_transform</span></code></a> avoid repeated
computation of the masked matrix. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/14356">#14356</a> by <a class="reference external" href="https://github.com/harsh020">Harsh Soni</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> <a class="reference internal" href="../modules/generated/sklearn.impute.IterativeImputer.html#sklearn.impute.IterativeImputer" title="sklearn.impute.IterativeImputer"><code class="xref py py-class docutils literal notranslate"><span class="pre">impute.IterativeImputer</span></code></a> now works when there is only one feature.
By <a class="reference external" href="https://github.com/sergeyf">Sergey Feldman</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> Fixed a bug in <a class="reference internal" href="../modules/generated/sklearn.impute.IterativeImputer.html#sklearn.impute.IterativeImputer" title="sklearn.impute.IterativeImputer"><code class="xref py py-class docutils literal notranslate"><span class="pre">impute.IterativeImputer</span></code></a> where features where
imputed in the reverse desired order with <code class="docutils literal notranslate"><span class="pre">imputation_order</span></code> either
<code class="docutils literal notranslate"><span class="pre">"ascending"</span></code> or <code class="docutils literal notranslate"><span class="pre">"descending"</span></code>. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/15393">#15393</a> by
<a class="reference external" href="https://github.com/venkyyuvy">Venkatachalam N</a>.</p></li>
</ul>
</div>
<div class="section" id="id7">
<h3><a class="reference internal" href="../modules/classes.html#module-sklearn.inspection" title="sklearn.inspection"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.inspection</span></code></a><a class="headerlink" href="#id7" title="Permalink to this headline">¶</a></h3>
<ul class="simple">
<li><p><span class="raw-html"><span class="badge badge-success">Major Feature</span></span> <a class="reference internal" href="../modules/generated/sklearn.inspection.permutation_importance.html#sklearn.inspection.permutation_importance" title="sklearn.inspection.permutation_importance"><code class="xref py py-func docutils literal notranslate"><span class="pre">inspection.permutation_importance</span></code></a> has been added to
measure the importance of each feature in an arbitrary trained model with
respect to a given scoring function. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/issues/13146">#13146</a> by <a class="reference external" href="https://github.com/thomasjpfan">Thomas Fan</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-success">Feature</span></span> <a class="reference internal" href="../modules/generated/sklearn.inspection.partial_dependence.html#sklearn.inspection.partial_dependence" title="sklearn.inspection.partial_dependence"><code class="xref py py-func docutils literal notranslate"><span class="pre">inspection.partial_dependence</span></code></a> and
<a class="reference internal" href="../modules/generated/sklearn.inspection.plot_partial_dependence.html#sklearn.inspection.plot_partial_dependence" title="sklearn.inspection.plot_partial_dependence"><code class="xref py py-func docutils literal notranslate"><span class="pre">inspection.plot_partial_dependence</span></code></a> now support the fast ‘recursion’
method for <a class="reference internal" href="../modules/generated/sklearn.ensemble.HistGradientBoostingClassifier.html#sklearn.ensemble.HistGradientBoostingClassifier" title="sklearn.ensemble.HistGradientBoostingClassifier"><code class="xref py py-class docutils literal notranslate"><span class="pre">ensemble.HistGradientBoostingClassifier</span></code></a> and
<a class="reference internal" href="../modules/generated/sklearn.ensemble.HistGradientBoostingRegressor.html#sklearn.ensemble.HistGradientBoostingRegressor" title="sklearn.ensemble.HistGradientBoostingRegressor"><code class="xref py py-class docutils literal notranslate"><span class="pre">ensemble.HistGradientBoostingRegressor</span></code></a>. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13769">#13769</a> by
<a class="reference external" href="https://github.com/NicolasHug">Nicolas Hug</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-info">Enhancement</span></span> <a class="reference internal" href="../modules/generated/sklearn.inspection.plot_partial_dependence.html#sklearn.inspection.plot_partial_dependence" title="sklearn.inspection.plot_partial_dependence"><code class="xref py py-func docutils literal notranslate"><span class="pre">inspection.plot_partial_dependence</span></code></a> has been extended to
now support the new visualization API described in the <a class="reference internal" href="../visualizations.html#visualizations"><span class="std std-ref">User Guide</span></a>. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/14646">#14646</a> by <a class="reference external" href="https://github.com/thomasjpfan">Thomas Fan</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-info">Enhancement</span></span> <a class="reference internal" href="../modules/generated/sklearn.inspection.partial_dependence.html#sklearn.inspection.partial_dependence" title="sklearn.inspection.partial_dependence"><code class="xref py py-func docutils literal notranslate"><span class="pre">inspection.partial_dependence</span></code></a> accepts pandas DataFrame
and <a class="reference internal" href="../modules/generated/sklearn.pipeline.Pipeline.html#sklearn.pipeline.Pipeline" title="sklearn.pipeline.Pipeline"><code class="xref py py-class docutils literal notranslate"><span class="pre">pipeline.Pipeline</span></code></a> containing <a class="reference internal" href="../modules/generated/sklearn.compose.ColumnTransformer.html#sklearn.compose.ColumnTransformer" title="sklearn.compose.ColumnTransformer"><code class="xref py py-class docutils literal notranslate"><span class="pre">compose.ColumnTransformer</span></code></a>.
In addition <a class="reference internal" href="../modules/generated/sklearn.inspection.plot_partial_dependence.html#sklearn.inspection.plot_partial_dependence" title="sklearn.inspection.plot_partial_dependence"><code class="xref py py-func docutils literal notranslate"><span class="pre">inspection.plot_partial_dependence</span></code></a> will use the column
names by default when a dataframe is passed.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/14028">#14028</a> and <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/15429">#15429</a> by <a class="reference external" href="https://github.com/glemaitre">Guillaume Lemaitre</a>.</p></li>
</ul>
</div>
<div class="section" id="sklearn-kernel-approximation">
<h3><a class="reference internal" href="../modules/classes.html#module-sklearn.kernel_approximation" title="sklearn.kernel_approximation"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.kernel_approximation</span></code></a><a class="headerlink" href="#sklearn-kernel-approximation" title="Permalink to this headline">¶</a></h3>
<ul class="simple">
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> Fixed a bug where <a class="reference internal" href="../modules/generated/sklearn.kernel_approximation.Nystroem.html#sklearn.kernel_approximation.Nystroem" title="sklearn.kernel_approximation.Nystroem"><code class="xref py py-class docutils literal notranslate"><span class="pre">kernel_approximation.Nystroem</span></code></a> raised a
<code class="docutils literal notranslate"><span class="pre">KeyError</span></code> when using <code class="docutils literal notranslate"><span class="pre">kernel="precomputed"</span></code>.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/14706">#14706</a> by <a class="reference external" href="https://github.com/venkyyuvy">Venkatachalam N</a>.</p></li>
</ul>
</div>
<div class="section" id="sklearn-linear-model">
<h3><a class="reference internal" href="../modules/classes.html#module-sklearn.linear_model" title="sklearn.linear_model"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.linear_model</span></code></a><a class="headerlink" href="#sklearn-linear-model" title="Permalink to this headline">¶</a></h3>
<ul class="simple">
<li><p><span class="raw-html"><span class="badge badge-info">Efficiency</span></span> The ‘liblinear’ logistic regression solver is now faster and
requires less memory.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/14108">#14108</a>, <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/14170">#14170</a>, <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/14296">#14296</a> by <a class="reference external" href="https://github.com/alexhenrie">Alex Henrie</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-info">Enhancement</span></span> <a class="reference internal" href="../modules/generated/sklearn.linear_model.BayesianRidge.html#sklearn.linear_model.BayesianRidge" title="sklearn.linear_model.BayesianRidge"><code class="xref py py-class docutils literal notranslate"><span class="pre">linear_model.BayesianRidge</span></code></a> now accepts hyperparameters
<code class="docutils literal notranslate"><span class="pre">alpha_init</span></code> and <code class="docutils literal notranslate"><span class="pre">lambda_init</span></code> which can be used to set the initial value
of the maximization procedure in <a class="reference internal" href="../glossary.html#term-fit"><span class="xref std std-term">fit</span></a>.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13618">#13618</a> by <a class="reference external" href="https://github.com/c56pony">Yoshihiro Uchida</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> <a class="reference internal" href="../modules/generated/sklearn.linear_model.Ridge.html#sklearn.linear_model.Ridge" title="sklearn.linear_model.Ridge"><code class="xref py py-class docutils literal notranslate"><span class="pre">linear_model.Ridge</span></code></a> now correctly fits an intercept when <code class="docutils literal notranslate"><span class="pre">X</span></code> is
sparse, <code class="docutils literal notranslate"><span class="pre">solver="auto"</span></code> and <code class="docutils literal notranslate"><span class="pre">fit_intercept=True</span></code>, because the default solver
in this configuration has changed to <code class="docutils literal notranslate"><span class="pre">sparse_cg</span></code>, which can fit an intercept
with sparse data. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13995">#13995</a> by <a class="reference external" href="https://github.com/jeromedockes">Jérôme Dockès</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> <a class="reference internal" href="../modules/generated/sklearn.linear_model.Ridge.html#sklearn.linear_model.Ridge" title="sklearn.linear_model.Ridge"><code class="xref py py-class docutils literal notranslate"><span class="pre">linear_model.Ridge</span></code></a> with <code class="docutils literal notranslate"><span class="pre">solver='sag'</span></code> now accepts F-ordered
and non-contiguous arrays and makes a conversion instead of failing.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/14458">#14458</a> by <a class="reference external" href="https://github.com/glemaitre">Guillaume Lemaitre</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> <a class="reference internal" href="../modules/generated/sklearn.linear_model.LassoCV.html#sklearn.linear_model.LassoCV" title="sklearn.linear_model.LassoCV"><code class="xref py py-class docutils literal notranslate"><span class="pre">linear_model.LassoCV</span></code></a> no longer forces <code class="docutils literal notranslate"><span class="pre">precompute=False</span></code>
when fitting the final model. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/14591">#14591</a> by <a class="reference external" href="https://amueller.github.io/">Andreas Müller</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> <a class="reference internal" href="../modules/generated/sklearn.linear_model.RidgeCV.html#sklearn.linear_model.RidgeCV" title="sklearn.linear_model.RidgeCV"><code class="xref py py-class docutils literal notranslate"><span class="pre">linear_model.RidgeCV</span></code></a> and <a class="reference internal" href="../modules/generated/sklearn.linear_model.RidgeClassifierCV.html#sklearn.linear_model.RidgeClassifierCV" title="sklearn.linear_model.RidgeClassifierCV"><code class="xref py py-class docutils literal notranslate"><span class="pre">linear_model.RidgeClassifierCV</span></code></a>
now correctly scores when <code class="docutils literal notranslate"><span class="pre">cv=None</span></code>.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/14864">#14864</a> by <a class="reference external" href="https://github.com/venkyyuvy">Venkatachalam N</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> Fixed a bug in <a class="reference internal" href="../modules/generated/sklearn.linear_model.LogisticRegressionCV.html#sklearn.linear_model.LogisticRegressionCV" title="sklearn.linear_model.LogisticRegressionCV"><code class="xref py py-class docutils literal notranslate"><span class="pre">linear_model.LogisticRegressionCV</span></code></a> where the
<code class="docutils literal notranslate"><span class="pre">scores_</span></code>, <code class="docutils literal notranslate"><span class="pre">n_iter_</span></code> and <code class="docutils literal notranslate"><span class="pre">coefs_paths_</span></code> attribute would have a wrong
ordering with <code class="docutils literal notranslate"><span class="pre">penalty='elastic-net'</span></code>. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/15044">#15044</a> by <a class="reference external" href="https://github.com/NicolasHug">Nicolas Hug</a></p></li>
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> <a class="reference internal" href="../modules/generated/sklearn.linear_model.MultiTaskLassoCV.html#sklearn.linear_model.MultiTaskLassoCV" title="sklearn.linear_model.MultiTaskLassoCV"><code class="xref py py-class docutils literal notranslate"><span class="pre">linear_model.MultiTaskLassoCV</span></code></a> and
<a class="reference internal" href="../modules/generated/sklearn.linear_model.MultiTaskElasticNetCV.html#sklearn.linear_model.MultiTaskElasticNetCV" title="sklearn.linear_model.MultiTaskElasticNetCV"><code class="xref py py-class docutils literal notranslate"><span class="pre">linear_model.MultiTaskElasticNetCV</span></code></a> with X of dtype int
and <code class="docutils literal notranslate"><span class="pre">fit_intercept=True</span></code>.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/15086">#15086</a> by <a class="reference external" href="https://github.com/agramfort">Alex Gramfort</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> The liblinear solver now supports <code class="docutils literal notranslate"><span class="pre">sample_weight</span></code>.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/15038">#15038</a> by <a class="reference external" href="https://github.com/glemaitre">Guillaume Lemaitre</a>.</p></li>
</ul>
</div>
<div class="section" id="sklearn-manifold">
<h3><a class="reference internal" href="../modules/classes.html#module-sklearn.manifold" title="sklearn.manifold"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.manifold</span></code></a><a class="headerlink" href="#sklearn-manifold" title="Permalink to this headline">¶</a></h3>
<ul class="simple">
<li><p><span class="raw-html"><span class="badge badge-success">Feature</span></span> <a class="reference internal" href="../modules/generated/sklearn.manifold.Isomap.html#sklearn.manifold.Isomap" title="sklearn.manifold.Isomap"><code class="xref py py-class docutils literal notranslate"><span class="pre">manifold.Isomap</span></code></a>, <a class="reference internal" href="../modules/generated/sklearn.manifold.TSNE.html#sklearn.manifold.TSNE" title="sklearn.manifold.TSNE"><code class="xref py py-class docutils literal notranslate"><span class="pre">manifold.TSNE</span></code></a>, and
<a class="reference internal" href="../modules/generated/sklearn.manifold.SpectralEmbedding.html#sklearn.manifold.SpectralEmbedding" title="sklearn.manifold.SpectralEmbedding"><code class="xref py py-class docutils literal notranslate"><span class="pre">manifold.SpectralEmbedding</span></code></a> now accept precomputed sparse
neighbors graph as input. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/issues/10482">#10482</a> by <a class="reference external" href="https://github.com/TomDLT">Tom Dupre la Tour</a> and
<a class="reference external" href="https://github.com/thechargedneutron">Kumar Ashutosh</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-success">Feature</span></span> Exposed the <code class="docutils literal notranslate"><span class="pre">n_jobs</span></code> parameter in <a class="reference internal" href="../modules/generated/sklearn.manifold.TSNE.html#sklearn.manifold.TSNE" title="sklearn.manifold.TSNE"><code class="xref py py-class docutils literal notranslate"><span class="pre">manifold.TSNE</span></code></a> for
multi-core calculation of the neighbors graph. This parameter has no
impact when <code class="docutils literal notranslate"><span class="pre">metric="precomputed"</span></code> or (<code class="docutils literal notranslate"><span class="pre">metric="euclidean"</span></code> and
<code class="docutils literal notranslate"><span class="pre">method="exact"</span></code>). <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/issues/15082">#15082</a> by <a class="reference external" href="https://github.com/rth">Roman Yurchak</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-info">Efficiency</span></span> Improved efficiency of <a class="reference internal" href="../modules/generated/sklearn.manifold.TSNE.html#sklearn.manifold.TSNE" title="sklearn.manifold.TSNE"><code class="xref py py-class docutils literal notranslate"><span class="pre">manifold.TSNE</span></code></a> when
<code class="docutils literal notranslate"><span class="pre">method="barnes-hut"</span></code> by computing the gradient in parallel.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13213">#13213</a> by <a class="reference external" href="https://github.com/tommoral">Thomas Moreau</a></p></li>
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> Fixed a bug where <a class="reference internal" href="../modules/generated/sklearn.manifold.spectral_embedding.html#sklearn.manifold.spectral_embedding" title="sklearn.manifold.spectral_embedding"><code class="xref py py-func docutils literal notranslate"><span class="pre">manifold.spectral_embedding</span></code></a> (and therefore
<a class="reference internal" href="../modules/generated/sklearn.manifold.SpectralEmbedding.html#sklearn.manifold.SpectralEmbedding" title="sklearn.manifold.SpectralEmbedding"><code class="xref py py-class docutils literal notranslate"><span class="pre">manifold.SpectralEmbedding</span></code></a> and <a class="reference internal" href="../modules/generated/sklearn.cluster.SpectralClustering.html#sklearn.cluster.SpectralClustering" title="sklearn.cluster.SpectralClustering"><code class="xref py py-class docutils literal notranslate"><span class="pre">cluster.SpectralClustering</span></code></a>)
computed wrong eigenvalues with <code class="docutils literal notranslate"><span class="pre">eigen_solver='amg'</span></code> when
<code class="docutils literal notranslate"><span class="pre">n_samples</span> <span class="pre"><</span> <span class="pre">5</span> <span class="pre">*</span> <span class="pre">n_components</span></code>. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/14647">#14647</a> by <a class="reference external" href="https://amueller.github.io/">Andreas Müller</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> Fixed a bug in <a class="reference internal" href="../modules/generated/sklearn.manifold.spectral_embedding.html#sklearn.manifold.spectral_embedding" title="sklearn.manifold.spectral_embedding"><code class="xref py py-func docutils literal notranslate"><span class="pre">manifold.spectral_embedding</span></code></a> used in
<a class="reference internal" href="../modules/generated/sklearn.manifold.SpectralEmbedding.html#sklearn.manifold.SpectralEmbedding" title="sklearn.manifold.SpectralEmbedding"><code class="xref py py-class docutils literal notranslate"><span class="pre">manifold.SpectralEmbedding</span></code></a> and <a class="reference internal" href="../modules/generated/sklearn.cluster.SpectralClustering.html#sklearn.cluster.SpectralClustering" title="sklearn.cluster.SpectralClustering"><code class="xref py py-class docutils literal notranslate"><span class="pre">cluster.SpectralClustering</span></code></a>
where <code class="docutils literal notranslate"><span class="pre">eigen_solver="amg"</span></code> would sometimes result in a LinAlgError.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/issues/13393">#13393</a> by <a class="reference external" href="https://github.com/lobpcg">Andrew Knyazev</a>
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13707">#13707</a> by <a class="reference external" href="https://github.com/whitews">Scott White</a></p></li>
<li><p><span class="raw-html"><span class="badge badge-warning">API Change</span></span> Deprecate <code class="docutils literal notranslate"><span class="pre">training_data_</span></code> unused attribute in
<a class="reference internal" href="../modules/generated/sklearn.manifold.Isomap.html#sklearn.manifold.Isomap" title="sklearn.manifold.Isomap"><code class="xref py py-class docutils literal notranslate"><span class="pre">manifold.Isomap</span></code></a>. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/issues/10482">#10482</a> by <a class="reference external" href="https://github.com/TomDLT">Tom Dupre la Tour</a>.</p></li>
</ul>
</div>
<div class="section" id="id8">
<h3><a class="reference internal" href="../modules/classes.html#module-sklearn.metrics" title="sklearn.metrics"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.metrics</span></code></a><a class="headerlink" href="#id8" title="Permalink to this headline">¶</a></h3>
<ul class="simple">
<li><p><span class="raw-html"><span class="badge badge-success">Major Feature</span></span> <a class="reference internal" href="../modules/generated/sklearn.metrics.plot_roc_curve.html#sklearn.metrics.plot_roc_curve" title="sklearn.metrics.plot_roc_curve"><code class="xref py py-func docutils literal notranslate"><span class="pre">metrics.plot_roc_curve</span></code></a> has been added to plot roc
curves. This function introduces the visualization API described in
the <a class="reference internal" href="../visualizations.html#visualizations"><span class="std std-ref">User Guide</span></a>. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/14357">#14357</a> by <a class="reference external" href="https://github.com/thomasjpfan">Thomas Fan</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-success">Feature</span></span> Added a new parameter <code class="docutils literal notranslate"><span class="pre">zero_division</span></code> to multiple classification
metrics: <code class="xref py py-func docutils literal notranslate"><span class="pre">precision_score</span></code>, <code class="xref py py-func docutils literal notranslate"><span class="pre">recall_score</span></code>, <code class="xref py py-func docutils literal notranslate"><span class="pre">f1_score</span></code>,
<code class="xref py py-func docutils literal notranslate"><span class="pre">fbeta_score</span></code>, <code class="xref py py-func docutils literal notranslate"><span class="pre">precision_recall_fscore_support</span></code>,
<code class="xref py py-func docutils literal notranslate"><span class="pre">classification_report</span></code>. This allows to set returned value for
ill-defined metrics.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/14900">#14900</a> by <a class="reference external" href="https://github.com/marctorrellas">Marc Torrellas Socastro</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-success">Feature</span></span> Added the <a class="reference internal" href="../modules/generated/sklearn.metrics.pairwise.nan_euclidean_distances.html#sklearn.metrics.pairwise.nan_euclidean_distances" title="sklearn.metrics.pairwise.nan_euclidean_distances"><code class="xref py py-func docutils literal notranslate"><span class="pre">metrics.pairwise.nan_euclidean_distances</span></code></a> metric,
which calculates euclidean distances in the presence of missing values.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/issues/12852">#12852</a> by <a class="reference external" href="https://github.com/ashimb9">Ashim Bhattarai</a> and <a class="reference external" href="https://github.com/thomasjpfan">Thomas Fan</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-success">Feature</span></span> New ranking metrics <a class="reference internal" href="../modules/generated/sklearn.metrics.ndcg_score.html#sklearn.metrics.ndcg_score" title="sklearn.metrics.ndcg_score"><code class="xref py py-func docutils literal notranslate"><span class="pre">metrics.ndcg_score</span></code></a> and
<a class="reference internal" href="../modules/generated/sklearn.metrics.dcg_score.html#sklearn.metrics.dcg_score" title="sklearn.metrics.dcg_score"><code class="xref py py-func docutils literal notranslate"><span class="pre">metrics.dcg_score</span></code></a> have been added to compute Discounted Cumulative
Gain and Normalized Discounted Cumulative Gain. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/9951">#9951</a> by <a class="reference external" href="https://github.com/jeromedockes">Jérôme
Dockès</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-success">Feature</span></span> <a class="reference internal" href="../modules/generated/sklearn.metrics.plot_precision_recall_curve.html#sklearn.metrics.plot_precision_recall_curve" title="sklearn.metrics.plot_precision_recall_curve"><code class="xref py py-func docutils literal notranslate"><span class="pre">metrics.plot_precision_recall_curve</span></code></a> has been added to plot
precision recall curves. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/14936">#14936</a> by <a class="reference external" href="https://github.com/thomasjpfan">Thomas Fan</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-success">Feature</span></span> <a class="reference internal" href="../modules/generated/sklearn.metrics.plot_confusion_matrix.html#sklearn.metrics.plot_confusion_matrix" title="sklearn.metrics.plot_confusion_matrix"><code class="xref py py-func docutils literal notranslate"><span class="pre">metrics.plot_confusion_matrix</span></code></a> has been added to plot
confusion matrices. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/15083">#15083</a> by <a class="reference external" href="https://github.com/thomasjpfan">Thomas Fan</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-success">Feature</span></span> Added multiclass support to <a class="reference internal" href="../modules/generated/sklearn.metrics.roc_auc_score.html#sklearn.metrics.roc_auc_score" title="sklearn.metrics.roc_auc_score"><code class="xref py py-func docutils literal notranslate"><span class="pre">metrics.roc_auc_score</span></code></a> with
corresponding scorers <code class="docutils literal notranslate"><span class="pre">'roc_auc_ovr'</span></code>, <code class="docutils literal notranslate"><span class="pre">'roc_auc_ovo'</span></code>,
<code class="docutils literal notranslate"><span class="pre">'roc_auc_ovr_weighted'</span></code>, and <code class="docutils literal notranslate"><span class="pre">'roc_auc_ovo_weighted'</span></code>.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/12789">#12789</a> and <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/15274">#15274</a> by
<a class="reference external" href="https://github.com/kathyxchen">Kathy Chen</a>, <a class="reference external" href="https://github.com/maskani-moh">Mohamed Maskani</a>, and
<a class="reference external" href="https://github.com/thomasjpfan">Thomas Fan</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-success">Feature</span></span> Add <a class="reference internal" href="../modules/generated/sklearn.metrics.mean_tweedie_deviance.html#sklearn.metrics.mean_tweedie_deviance" title="sklearn.metrics.mean_tweedie_deviance"><code class="xref py py-class docutils literal notranslate"><span class="pre">metrics.mean_tweedie_deviance</span></code></a> measuring the
Tweedie deviance for a given <code class="docutils literal notranslate"><span class="pre">power</span></code> parameter. Also add mean Poisson
deviance <a class="reference internal" href="../modules/generated/sklearn.metrics.mean_poisson_deviance.html#sklearn.metrics.mean_poisson_deviance" title="sklearn.metrics.mean_poisson_deviance"><code class="xref py py-class docutils literal notranslate"><span class="pre">metrics.mean_poisson_deviance</span></code></a> and mean Gamma deviance
<a class="reference internal" href="../modules/generated/sklearn.metrics.mean_gamma_deviance.html#sklearn.metrics.mean_gamma_deviance" title="sklearn.metrics.mean_gamma_deviance"><code class="xref py py-class docutils literal notranslate"><span class="pre">metrics.mean_gamma_deviance</span></code></a> that are special cases of the Tweedie
deviance for <code class="docutils literal notranslate"><span class="pre">power=1</span></code> and <code class="docutils literal notranslate"><span class="pre">power=2</span></code> respectively.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13938">#13938</a> by <a class="reference external" href="https://github.com/lorentzenchr">Christian Lorentzen</a> and
<a class="reference external" href="https://github.com/rth">Roman Yurchak</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-info">Efficiency</span></span> Improved performance of
<a class="reference internal" href="../modules/generated/sklearn.metrics.pairwise.manhattan_distances.html#sklearn.metrics.pairwise.manhattan_distances" title="sklearn.metrics.pairwise.manhattan_distances"><code class="xref py py-func docutils literal notranslate"><span class="pre">metrics.pairwise.manhattan_distances</span></code></a> in the case of sparse matrices.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/15049">#15049</a> by <code class="docutils literal notranslate"><span class="pre">Paolo</span> <span class="pre">Toccaceli</span> <span class="pre"><ptocca></span></code>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-info">Enhancement</span></span> The parameter <code class="docutils literal notranslate"><span class="pre">beta</span></code> in <a class="reference internal" href="../modules/generated/sklearn.metrics.fbeta_score.html#sklearn.metrics.fbeta_score" title="sklearn.metrics.fbeta_score"><code class="xref py py-func docutils literal notranslate"><span class="pre">metrics.fbeta_score</span></code></a> is
updated to accept the zero and <code class="docutils literal notranslate"><span class="pre">float('+inf')</span></code> value.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13231">#13231</a> by <a class="reference external" href="https://github.com/corona10">Dong-hee Na</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-info">Enhancement</span></span> Added parameter <code class="docutils literal notranslate"><span class="pre">squared</span></code> in <a class="reference internal" href="../modules/generated/sklearn.metrics.mean_squared_error.html#sklearn.metrics.mean_squared_error" title="sklearn.metrics.mean_squared_error"><code class="xref py py-func docutils literal notranslate"><span class="pre">metrics.mean_squared_error</span></code></a>
to return root mean squared error.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13467">#13467</a> by <a class="reference external" href="https://github.com/urvang96">Urvang Patel</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-info">Enhancement</span></span> Allow computing averaged metrics in the case of no true positives.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/14595">#14595</a> by <a class="reference external" href="https://amueller.github.io/">Andreas Müller</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-info">Enhancement</span></span> Multilabel metrics now supports list of lists as input.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/14865">#14865</a> <a class="reference external" href="https://github.com/srivatsan-ramesh">Srivatsan Ramesh</a>,
<a class="reference external" href="https://github.com/herilalaina">Herilalaina Rakotoarison</a>,
<a class="reference external" href="https://github.com/leonardbinet">Léonard Binet</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-info">Enhancement</span></span> <a class="reference internal" href="../modules/generated/sklearn.metrics.median_absolute_error.html#sklearn.metrics.median_absolute_error" title="sklearn.metrics.median_absolute_error"><code class="xref py py-func docutils literal notranslate"><span class="pre">metrics.median_absolute_error</span></code></a> now supports
<code class="docutils literal notranslate"><span class="pre">multioutput</span></code> parameter.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/14732">#14732</a> by <a class="reference external" href="https://github.com/agamemnonc">Agamemnon Krasoulis</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-info">Enhancement</span></span> ‘roc_auc_ovr_weighted’ and ‘roc_auc_ovo_weighted’ can now be
used as the <a class="reference internal" href="../glossary.html#term-scoring"><span class="xref std std-term">scoring</span></a> parameter of model-selection tools.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/14417">#14417</a> by <a class="reference external" href="https://github.com/thomasjpfan">Thomas Fan</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-info">Enhancement</span></span> <a class="reference internal" href="../modules/generated/sklearn.metrics.confusion_matrix.html#sklearn.metrics.confusion_matrix" title="sklearn.metrics.confusion_matrix"><code class="xref py py-func docutils literal notranslate"><span class="pre">metrics.confusion_matrix</span></code></a> accepts a parameters
<code class="docutils literal notranslate"><span class="pre">normalize</span></code> allowing to normalize the confusion matrix by column, rows, or
overall.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/15625">#15625</a> by <code class="docutils literal notranslate"><span class="pre">Guillaume</span> <span class="pre">Lemaitre</span> <span class="pre"><glemaitre></span></code>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> Raise a ValueError in <a class="reference internal" href="../modules/generated/sklearn.metrics.silhouette_score.html#sklearn.metrics.silhouette_score" title="sklearn.metrics.silhouette_score"><code class="xref py py-func docutils literal notranslate"><span class="pre">metrics.silhouette_score</span></code></a> when a
precomputed distance matrix contains non-zero diagonal entries.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/12258">#12258</a> by <a class="reference external" href="https://github.com/sjtrny">Stephen Tierney</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-warning">API Change</span></span> <code class="docutils literal notranslate"><span class="pre">scoring="neg_brier_score"</span></code> should be used instead of
<code class="docutils literal notranslate"><span class="pre">scoring="brier_score_loss"</span></code> which is now deprecated.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/14898">#14898</a> by <a class="reference external" href="https://github.com/stefan-matcovici">Stefan Matcovici</a>.</p></li>
</ul>
</div>
<div class="section" id="id9">
<h3><a class="reference internal" href="../modules/classes.html#module-sklearn.model_selection" title="sklearn.model_selection"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.model_selection</span></code></a><a class="headerlink" href="#id9" title="Permalink to this headline">¶</a></h3>
<ul class="simple">
<li><p><span class="raw-html"><span class="badge badge-info">Efficiency</span></span> Improved performance of multimetric scoring in
<a class="reference internal" href="../modules/generated/sklearn.model_selection.cross_validate.html#sklearn.model_selection.cross_validate" title="sklearn.model_selection.cross_validate"><code class="xref py py-func docutils literal notranslate"><span class="pre">model_selection.cross_validate</span></code></a>,
<a class="reference internal" href="../modules/generated/sklearn.model_selection.GridSearchCV.html#sklearn.model_selection.GridSearchCV" title="sklearn.model_selection.GridSearchCV"><code class="xref py py-class docutils literal notranslate"><span class="pre">model_selection.GridSearchCV</span></code></a>, and
<a class="reference internal" href="../modules/generated/sklearn.model_selection.RandomizedSearchCV.html#sklearn.model_selection.RandomizedSearchCV" title="sklearn.model_selection.RandomizedSearchCV"><code class="xref py py-class docutils literal notranslate"><span class="pre">model_selection.RandomizedSearchCV</span></code></a>. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/14593">#14593</a> by <a class="reference external" href="https://github.com/thomasjpfan">Thomas Fan</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-info">Enhancement</span></span> <a class="reference internal" href="../modules/generated/sklearn.model_selection.learning_curve.html#sklearn.model_selection.learning_curve" title="sklearn.model_selection.learning_curve"><code class="xref py py-class docutils literal notranslate"><span class="pre">model_selection.learning_curve</span></code></a> now accepts parameter
<code class="docutils literal notranslate"><span class="pre">return_times</span></code> which can be used to retrieve computation times in order to
plot model scalability (see learning_curve example).
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13938">#13938</a> by <a class="reference external" href="https://github.com/H4dr1en">Hadrien Reboul</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-info">Enhancement</span></span> <a class="reference internal" href="../modules/generated/sklearn.model_selection.RandomizedSearchCV.html#sklearn.model_selection.RandomizedSearchCV" title="sklearn.model_selection.RandomizedSearchCV"><code class="xref py py-class docutils literal notranslate"><span class="pre">model_selection.RandomizedSearchCV</span></code></a> now accepts lists
of parameter distributions. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/14549">#14549</a> by <a class="reference external" href="https://amueller.github.io/">Andreas Müller</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> Reimplemented <a class="reference internal" href="../modules/generated/sklearn.model_selection.StratifiedKFold.html#sklearn.model_selection.StratifiedKFold" title="sklearn.model_selection.StratifiedKFold"><code class="xref py py-class docutils literal notranslate"><span class="pre">model_selection.StratifiedKFold</span></code></a> to fix an issue
where one test set could be <code class="docutils literal notranslate"><span class="pre">n_classes</span></code> larger than another. Test sets should
now be near-equally sized. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/14704">#14704</a> by <a class="reference external" href="https://joelnothman.com/">Joel Nothman</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> The <code class="docutils literal notranslate"><span class="pre">cv_results_</span></code> attribute of <a class="reference internal" href="../modules/generated/sklearn.model_selection.GridSearchCV.html#sklearn.model_selection.GridSearchCV" title="sklearn.model_selection.GridSearchCV"><code class="xref py py-class docutils literal notranslate"><span class="pre">model_selection.GridSearchCV</span></code></a>
and <a class="reference internal" href="../modules/generated/sklearn.model_selection.RandomizedSearchCV.html#sklearn.model_selection.RandomizedSearchCV" title="sklearn.model_selection.RandomizedSearchCV"><code class="xref py py-class docutils literal notranslate"><span class="pre">model_selection.RandomizedSearchCV</span></code></a> now only contains unfitted
estimators. This potentially saves a lot of memory since the state of the
estimators isn’t stored. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/#15096">##15096</a> by <a class="reference external" href="https://amueller.github.io/">Andreas Müller</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-warning">API Change</span></span> <a class="reference internal" href="../modules/generated/sklearn.model_selection.KFold.html#sklearn.model_selection.KFold" title="sklearn.model_selection.KFold"><code class="xref py py-class docutils literal notranslate"><span class="pre">model_selection.KFold</span></code></a> and
<a class="reference internal" href="../modules/generated/sklearn.model_selection.StratifiedKFold.html#sklearn.model_selection.StratifiedKFold" title="sklearn.model_selection.StratifiedKFold"><code class="xref py py-class docutils literal notranslate"><span class="pre">model_selection.StratifiedKFold</span></code></a> now raise a warning if
<code class="docutils literal notranslate"><span class="pre">random_state</span></code> is set but <code class="docutils literal notranslate"><span class="pre">shuffle</span></code> is False. This will raise an error in
0.24.</p></li>
</ul>
</div>
<div class="section" id="sklearn-multioutput">
<h3><a class="reference internal" href="../modules/classes.html#module-sklearn.multioutput" title="sklearn.multioutput"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.multioutput</span></code></a><a class="headerlink" href="#sklearn-multioutput" title="Permalink to this headline">¶</a></h3>
<ul class="simple">
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> <a class="reference internal" href="../modules/generated/sklearn.multioutput.MultiOutputClassifier.html#sklearn.multioutput.MultiOutputClassifier" title="sklearn.multioutput.MultiOutputClassifier"><code class="xref py py-class docutils literal notranslate"><span class="pre">multioutput.MultiOutputClassifier</span></code></a> now has attribute
<code class="docutils literal notranslate"><span class="pre">classes_</span></code>. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/14629">#14629</a> by <a class="reference external" href="https://github.com/agamemnonc">Agamemnon Krasoulis</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> <a class="reference internal" href="../modules/generated/sklearn.multioutput.MultiOutputClassifier.html#sklearn.multioutput.MultiOutputClassifier" title="sklearn.multioutput.MultiOutputClassifier"><code class="xref py py-class docutils literal notranslate"><span class="pre">multioutput.MultiOutputClassifier</span></code></a> now has <code class="docutils literal notranslate"><span class="pre">predict_proba</span></code>
as property and can be checked with <code class="docutils literal notranslate"><span class="pre">hasattr</span></code>.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/issues/15488">#15488</a> <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/15490">#15490</a> by <a class="reference external" href="https://github.com/rebekahkim">Rebekah Kim</a></p></li>
</ul>
</div>
<div class="section" id="id10">
<h3><a class="reference internal" href="../modules/classes.html#module-sklearn.naive_bayes" title="sklearn.naive_bayes"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.naive_bayes</span></code></a><a class="headerlink" href="#id10" title="Permalink to this headline">¶</a></h3>
<ul class="simple">
<li><p><span class="raw-html"><span class="badge badge-success">Major Feature</span></span> Added <a class="reference internal" href="../modules/generated/sklearn.naive_bayes.CategoricalNB.html#sklearn.naive_bayes.CategoricalNB" title="sklearn.naive_bayes.CategoricalNB"><code class="xref py py-class docutils literal notranslate"><span class="pre">naive_bayes.CategoricalNB</span></code></a> that implements the
Categorical Naive Bayes classifier.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/12569">#12569</a> by <a class="reference external" href="https://github.com/timbicker">Tim Bicker</a> and
<a class="reference external" href="https://github.com/FlorianWilhelm">Florian Wilhelm</a>.</p></li>
</ul>
</div>
<div class="section" id="id11">
<h3><a class="reference internal" href="../modules/classes.html#module-sklearn.neighbors" title="sklearn.neighbors"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.neighbors</span></code></a><a class="headerlink" href="#id11" title="Permalink to this headline">¶</a></h3>
<ul class="simple">
<li><p><span class="raw-html"><span class="badge badge-success">Major Feature</span></span> Added <a class="reference internal" href="../modules/generated/sklearn.neighbors.KNeighborsTransformer.html#sklearn.neighbors.KNeighborsTransformer" title="sklearn.neighbors.KNeighborsTransformer"><code class="xref py py-class docutils literal notranslate"><span class="pre">neighbors.KNeighborsTransformer</span></code></a> and
<a class="reference internal" href="../modules/generated/sklearn.neighbors.RadiusNeighborsTransformer.html#sklearn.neighbors.RadiusNeighborsTransformer" title="sklearn.neighbors.RadiusNeighborsTransformer"><code class="xref py py-class docutils literal notranslate"><span class="pre">neighbors.RadiusNeighborsTransformer</span></code></a>, which transform input dataset
into a sparse neighbors graph. They give finer control on nearest neighbors
computations and enable easy pipeline caching for multiple use.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/issues/10482">#10482</a> by <a class="reference external" href="https://github.com/TomDLT">Tom Dupre la Tour</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-success">Feature</span></span> <a class="reference internal" href="../modules/generated/sklearn.neighbors.KNeighborsClassifier.html#sklearn.neighbors.KNeighborsClassifier" title="sklearn.neighbors.KNeighborsClassifier"><code class="xref py py-class docutils literal notranslate"><span class="pre">neighbors.KNeighborsClassifier</span></code></a>,
<a class="reference internal" href="../modules/generated/sklearn.neighbors.KNeighborsRegressor.html#sklearn.neighbors.KNeighborsRegressor" title="sklearn.neighbors.KNeighborsRegressor"><code class="xref py py-class docutils literal notranslate"><span class="pre">neighbors.KNeighborsRegressor</span></code></a>,
<a class="reference internal" href="../modules/generated/sklearn.neighbors.RadiusNeighborsClassifier.html#sklearn.neighbors.RadiusNeighborsClassifier" title="sklearn.neighbors.RadiusNeighborsClassifier"><code class="xref py py-class docutils literal notranslate"><span class="pre">neighbors.RadiusNeighborsClassifier</span></code></a>,
<a class="reference internal" href="../modules/generated/sklearn.neighbors.RadiusNeighborsRegressor.html#sklearn.neighbors.RadiusNeighborsRegressor" title="sklearn.neighbors.RadiusNeighborsRegressor"><code class="xref py py-class docutils literal notranslate"><span class="pre">neighbors.RadiusNeighborsRegressor</span></code></a>, and
<a class="reference internal" href="../modules/generated/sklearn.neighbors.LocalOutlierFactor.html#sklearn.neighbors.LocalOutlierFactor" title="sklearn.neighbors.LocalOutlierFactor"><code class="xref py py-class docutils literal notranslate"><span class="pre">neighbors.LocalOutlierFactor</span></code></a> now accept precomputed sparse
neighbors graph as input. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/issues/10482">#10482</a> by <a class="reference external" href="https://github.com/TomDLT">Tom Dupre la Tour</a> and
<a class="reference external" href="https://github.com/thechargedneutron">Kumar Ashutosh</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-success">Feature</span></span> <a class="reference internal" href="../modules/generated/sklearn.neighbors.RadiusNeighborsClassifier.html#sklearn.neighbors.RadiusNeighborsClassifier" title="sklearn.neighbors.RadiusNeighborsClassifier"><code class="xref py py-class docutils literal notranslate"><span class="pre">neighbors.RadiusNeighborsClassifier</span></code></a> now supports
predicting probabilities by using <code class="docutils literal notranslate"><span class="pre">predict_proba</span></code> and supports more