-
Notifications
You must be signed in to change notification settings - Fork 81
/
Copy pathv0.21.html
1385 lines (1351 loc) · 197 KB
/
v0.21.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="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
<meta property="og:title" content="Version 0.21.3" />
<meta property="og:type" content="website" />
<meta property="og:url" content="https://scikit-learn/stable/whats_new/v0.21.html" />
<meta property="og:site_name" content="scikit-learn" />
<meta property="og:description" content="Legend for changelogs:: something big that you couldn’t do before.,: something that you couldn’t do before.,: an existing feature now may not require as much computation or memory.,: a miscellaneou..." />
<meta property="og:image" content="https://scikit-learn.org/stable/_static/scikit-learn-logo-small.png" />
<meta property="og:image:alt" content="scikit-learn" />
<meta name="description" content="Legend for changelogs:: something big that you couldn’t do before.,: something that you couldn’t do before.,: an existing feature now may not require as much computation or memory.,: a miscellaneou..." />
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<title>Version 0.21.3 — scikit-learn 1.2.2 documentation</title>
<link rel="canonical" href="http://scikit-learn.org/stable/whats_new/v0.21.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/pygments.css" type="text/css" />
<link rel="stylesheet" href="../_static/css/theme.css" type="text/css" />
<link rel="stylesheet" href="../_static/plot_directive.css" type="text/css" />
<link rel="stylesheet" href="../_static/sg_gallery.css" type="text/css" />
<link rel="stylesheet" href="../_static/sg_gallery-binder.css" type="text/css" />
<link rel="stylesheet" href="../_static/sg_gallery-dataframe.css" type="text/css" />
<link rel="stylesheet" href="../_static/sg_gallery-rendered-html.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/js/vendor/jquery-3.6.3.slim.min.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" target="_blank" rel="noopener noreferrer" href="https://blog.scikit-learn.org/">Community</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="v1.2.html" >What's new</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="https://scikit-learn.org/dev/developers/index.html" target="_blank" rel="noopener noreferrer">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="../support.html" >Support</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="../governance.html" >Governance</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 and Download</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="v1.2.html" >What's new</a>
<a class="sk-nav-dropdown-item dropdown-item" href="../glossary.html" >Glossary</a>
<a class="sk-nav-dropdown-item dropdown-item" href="https://scikit-learn.org/dev/developers/index.html" target="_blank" rel="noopener noreferrer">Development</a>
<a class="sk-nav-dropdown-item dropdown-item" href="../faq.html" >FAQ</a>
<a class="sk-nav-dropdown-item dropdown-item" href="../support.html" >Support</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="../governance.html" >Governance</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 and Download</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="btn-group w-100 mb-2" role="group" aria-label="rellinks">
<a href="v0.22.html" role="button" class="btn sk-btn-rellink py-1" sk-rellink-tooltip="Version 0.22.2.post1">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="v0.20.html" role="button" class="btn sk-btn-rellink py-1" sk-rellink-tooltip="Version 0.20.4">Next</a>
</div>
<div class="alert alert-danger p-1 mb-2" role="alert">
<p class="text-center mb-0">
<strong>scikit-learn 1.2.2</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 0.21.3</a><ul>
<li><a class="reference internal" href="#legend-for-changelogs">Legend for changelogs</a></li>
<li><a class="reference internal" href="#changed-models">Changed models</a></li>
<li><a class="reference internal" href="#changelog">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-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-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-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-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-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="#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-neighbors"><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-tree"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.tree</span></code></a></li>
</ul>
</li>
</ul>
</li>
<li><a class="reference internal" href="#version-0-21-2">Version 0.21.2</a><ul>
<li><a class="reference internal" href="#id1">Changelog</a><ul>
<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-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-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-utils-sparsefuncs"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.utils.sparsefuncs</span></code></a></li>
</ul>
</li>
</ul>
</li>
<li><a class="reference internal" href="#version-0-21-1">Version 0.21.1</a><ul>
<li><a class="reference internal" href="#id2">Changelog</a><ul>
<li><a class="reference internal" href="#id3"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.inspection</span></code></a></li>
<li><a class="reference internal" href="#id4"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.metrics</span></code></a></li>
<li><a class="reference internal" href="#id5"><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-0-21-0">Version 0.21.0</a><ul>
<li><a class="reference internal" href="#id6">Changed models</a></li>
<li><a class="reference internal" href="#known-major-bugs">Known Major Bugs</a></li>
<li><a class="reference internal" href="#id7">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="#id8"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.cluster</span></code></a></li>
<li><a class="reference internal" href="#id9"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.compose</span></code></a></li>
<li><a class="reference internal" href="#id10"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.datasets</span></code></a></li>
<li><a class="reference internal" href="#id11"><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-discriminant-analysis"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.discriminant_analysis</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="#id12"><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-externals"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.externals</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="#id13"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.impute</span></code></a></li>
<li><a class="reference internal" href="#id14"><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-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="#id15"><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="#id16"><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-mixture"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.mixture</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-multiclass"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.multiclass</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="#id17"><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="#id18"><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-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="#id19"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.tree</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>
<li><a class="reference internal" href="#multiple-modules">Multiple modules</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">
<section id="version-0-21-3">
<span id="changes-0-21-3"></span><h1>Version 0.21.3<a class="headerlink" href="#version-0-21-3" title="Permalink to this heading">¶</a></h1>
<section id="legend-for-changelogs">
<h2>Legend for changelogs<a class="headerlink" href="#legend-for-changelogs" title="Permalink to this heading">¶</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>
<p><strong>July 30, 2019</strong></p>
</section>
<section id="changed-models">
<h2>Changed models<a class="headerlink" href="#changed-models" title="Permalink to this heading">¶</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>The v0.20.0 release notes failed to mention a backwards incompatibility in
<a class="reference internal" href="../modules/generated/sklearn.metrics.make_scorer.html#sklearn.metrics.make_scorer" title="sklearn.metrics.make_scorer"><code class="xref py py-func docutils literal notranslate"><span class="pre">metrics.make_scorer</span></code></a> when <code class="docutils literal notranslate"><span class="pre">needs_proba=True</span></code> and <code class="docutils literal notranslate"><span class="pre">y_true</span></code> is binary.
Now, the scorer function is supposed to accept a 1D <code class="docutils literal notranslate"><span class="pre">y_pred</span></code> (i.e.,
probability of the positive class, shape <code class="docutils literal notranslate"><span class="pre">(n_samples,)</span></code>), instead of a 2D
<code class="docutils literal notranslate"><span class="pre">y_pred</span></code> (i.e., shape <code class="docutils literal notranslate"><span class="pre">(n_samples,</span> <span class="pre">2)</span></code>).</p></li>
</ul>
</section>
<section id="changelog">
<h2>Changelog<a class="headerlink" href="#changelog" title="Permalink to this heading">¶</a></h2>
<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 heading">¶</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.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> where computation with
<code class="docutils literal notranslate"><span class="pre">init='random'</span></code> was single threaded for <code class="docutils literal notranslate"><span class="pre">n_jobs</span> <span class="pre">></span> <span class="pre">1</span></code> or <code class="docutils literal notranslate"><span class="pre">n_jobs</span> <span class="pre">=</span> <span class="pre">-1</span></code>.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/12955">#12955</a> by <a class="reference external" href="https://github.com/nixphix">Prabakaran Kumaresshan</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.cluster.OPTICS.html#sklearn.cluster.OPTICS" title="sklearn.cluster.OPTICS"><code class="xref py py-class docutils literal notranslate"><span class="pre">cluster.OPTICS</span></code></a> where users were unable to pass
float <code class="docutils literal notranslate"><span class="pre">min_samples</span></code> and <code class="docutils literal notranslate"><span class="pre">min_cluster_size</span></code>. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/14496">#14496</a> by
<a class="reference external" href="https://github.com/someusername1">Fabian Klopfer</a>
and <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> Fixed a bug 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> where KMeans++ initialisation
could rarely result in an IndexError. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/issues/11756">#11756</a> by <a class="reference external" href="https://joelnothman.com/">Joel Nothman</a>.</p></li>
</ul>
</section>
<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 heading">¶</a></h3>
<ul class="simple">
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> Fixed an issue 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> where using
DataFrames whose column order differs between :func:<code class="docutils literal notranslate"><span class="pre">fit</span></code> and
:func:<code class="docutils literal notranslate"><span class="pre">transform</span></code> could lead to silently passing incorrect columns to the
<code class="docutils literal notranslate"><span class="pre">remainder</span></code> transformer.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/14237">#14237</a> by <code class="docutils literal notranslate"><span class="pre">Andreas</span> <span class="pre">Schuderer</span> <span class="pre"><schuderer></span></code>.</p></li>
</ul>
</section>
<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 heading">¶</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.datasets.fetch_california_housing.html#sklearn.datasets.fetch_california_housing" title="sklearn.datasets.fetch_california_housing"><code class="xref py py-func docutils literal notranslate"><span class="pre">datasets.fetch_california_housing</span></code></a>,
<a class="reference internal" href="../modules/generated/sklearn.datasets.fetch_covtype.html#sklearn.datasets.fetch_covtype" title="sklearn.datasets.fetch_covtype"><code class="xref py py-func docutils literal notranslate"><span class="pre">datasets.fetch_covtype</span></code></a>,
<a class="reference internal" href="../modules/generated/sklearn.datasets.fetch_kddcup99.html#sklearn.datasets.fetch_kddcup99" title="sklearn.datasets.fetch_kddcup99"><code class="xref py py-func docutils literal notranslate"><span class="pre">datasets.fetch_kddcup99</span></code></a>, <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 internal" href="../modules/generated/sklearn.datasets.fetch_rcv1.html#sklearn.datasets.fetch_rcv1" title="sklearn.datasets.fetch_rcv1"><code class="xref py py-func docutils literal notranslate"><span class="pre">datasets.fetch_rcv1</span></code></a>, and <a class="reference internal" href="../modules/generated/sklearn.datasets.fetch_species_distributions.html#sklearn.datasets.fetch_species_distributions" title="sklearn.datasets.fetch_species_distributions"><code class="xref py py-func docutils literal notranslate"><span class="pre">datasets.fetch_species_distributions</span></code></a>
try to persist the previously cache using the new <code class="docutils literal notranslate"><span class="pre">joblib</span></code> if the cached
data was persisted using the deprecated <code class="docutils literal notranslate"><span class="pre">sklearn.externals.joblib</span></code>. This
behavior is set to be deprecated and removed in v0.23.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/14197">#14197</a> by <a class="reference external" href="https://github.com/adrinjalali">Adrin Jalali</a>.</p></li>
</ul>
</section>
<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 heading">¶</a></h3>
<ul class="simple">
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> Fix zero division error in <code class="xref py py-func docutils literal notranslate"><span class="pre">HistGradientBoostingClassifier</span></code> and
<code class="xref py py-func docutils literal notranslate"><span class="pre">HistGradientBoostingRegressor</span></code>.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/14024">#14024</a> by <code class="docutils literal notranslate"><span class="pre">Nicolas</span> <span class="pre">Hug</span> <span class="pre"><NicolasHug></span></code>.</p></li>
</ul>
</section>
<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 heading">¶</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.impute.SimpleImputer.html#sklearn.impute.SimpleImputer" title="sklearn.impute.SimpleImputer"><code class="xref py py-class docutils literal notranslate"><span class="pre">impute.SimpleImputer</span></code></a> and
<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> so that no errors are thrown when there are
missing values in training data. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13974">#13974</a> by <code class="docutils literal notranslate"><span class="pre">Frank</span> <span class="pre">Hoang</span> <span class="pre"><fhoang7></span></code>.</p></li>
</ul>
</section>
<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 heading">¶</a></h3>
<ul class="simple">
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> Fixed a bug in <code class="xref py py-func docutils literal notranslate"><span class="pre">inspection.plot_partial_dependence</span></code> where
<code class="docutils literal notranslate"><span class="pre">target</span></code> parameter was not being taken into account for multiclass problems.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/14393">#14393</a> by <a class="reference external" href="https://github.com/guillemgsubies">Guillem G. Subies</a>.</p></li>
</ul>
</section>
<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 heading">¶</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.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
<code class="docutils literal notranslate"><span class="pre">refit=False</span></code> would fail depending on the <code class="docutils literal notranslate"><span class="pre">'multiclass'</span></code> and
<code class="docutils literal notranslate"><span class="pre">'penalty'</span></code> parameters (regression introduced in 0.21). <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/14087">#14087</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> Compatibility fix for <a class="reference internal" href="../modules/generated/sklearn.linear_model.ARDRegression.html#sklearn.linear_model.ARDRegression" title="sklearn.linear_model.ARDRegression"><code class="xref py py-class docutils literal notranslate"><span class="pre">linear_model.ARDRegression</span></code></a> and
Scipy>=1.3.0. Adapts to upstream changes to the default <code class="docutils literal notranslate"><span class="pre">pinvh</span></code> cutoff
threshold which otherwise results in poor accuracy in some cases.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/14067">#14067</a> by <a class="reference external" href="https://github.com/timstaley">Tim Staley</a>.</p></li>
</ul>
</section>
<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 heading">¶</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.neighbors.NeighborhoodComponentsAnalysis.html#sklearn.neighbors.NeighborhoodComponentsAnalysis" title="sklearn.neighbors.NeighborhoodComponentsAnalysis"><code class="xref py py-class docutils literal notranslate"><span class="pre">neighbors.NeighborhoodComponentsAnalysis</span></code></a> where
the validation of initial parameters <code class="docutils literal notranslate"><span class="pre">n_components</span></code>, <code class="docutils literal notranslate"><span class="pre">max_iter</span></code> and
<code class="docutils literal notranslate"><span class="pre">tol</span></code> required too strict types. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/14092">#14092</a> by
<a class="reference external" href="https://github.com/jeremiedbb">Jérémie du Boisberranger</a>.</p></li>
</ul>
</section>
<section id="sklearn-tree">
<h3><a class="reference internal" href="../modules/classes.html#module-sklearn.tree" title="sklearn.tree"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.tree</span></code></a><a class="headerlink" href="#sklearn-tree" title="Permalink to this heading">¶</a></h3>
<ul class="simple">
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> Fixed bug in <a class="reference internal" href="../modules/generated/sklearn.tree.export_text.html#sklearn.tree.export_text" title="sklearn.tree.export_text"><code class="xref py py-func docutils literal notranslate"><span class="pre">tree.export_text</span></code></a> when the tree has one feature and
a single feature name is passed in. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/14053">#14053</a> by <code class="docutils literal notranslate"><span class="pre">Thomas</span> <span class="pre">Fan</span></code>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> Fixed an issue with <code class="xref py py-func docutils literal notranslate"><span class="pre">plot_tree</span></code> where it displayed
entropy calculations even for <code class="docutils literal notranslate"><span class="pre">gini</span></code> criterion in DecisionTreeClassifiers.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13947">#13947</a> by <a class="reference external" href="https://github.com/fhoang7">Frank Hoang</a>.</p></li>
</ul>
</section>
</section>
</section>
<section id="version-0-21-2">
<span id="changes-0-21-2"></span><h1>Version 0.21.2<a class="headerlink" href="#version-0-21-2" title="Permalink to this heading">¶</a></h1>
<p><strong>24 May 2019</strong></p>
<section id="id1">
<h2>Changelog<a class="headerlink" href="#id1" title="Permalink to this heading">¶</a></h2>
<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 heading">¶</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.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> improving numerical
stability when <code class="docutils literal notranslate"><span class="pre">Y</span></code> is close to zero. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13903">#13903</a> by <a class="reference external" href="https://github.com/thomasjpfan">Thomas Fan</a>.</p></li>
</ul>
</section>
<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 heading">¶</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.pairwise.euclidean_distances.html#sklearn.metrics.pairwise.euclidean_distances" title="sklearn.metrics.pairwise.euclidean_distances"><code class="xref py py-func docutils literal notranslate"><span class="pre">metrics.pairwise.euclidean_distances</span></code></a> where a
part of the distance matrix was left un-instanciated for suffiently large
float32 datasets (regression introduced in 0.21). <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13910">#13910</a> by
<a class="reference external" href="https://github.com/jeremiedbb">Jérémie du Boisberranger</a>.</p></li>
</ul>
</section>
<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 heading">¶</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.preprocessing.OneHotEncoder.html#sklearn.preprocessing.OneHotEncoder" title="sklearn.preprocessing.OneHotEncoder"><code class="xref py py-class docutils literal notranslate"><span class="pre">preprocessing.OneHotEncoder</span></code></a> where the new
<code class="docutils literal notranslate"><span class="pre">drop</span></code> parameter was not reflected in <code class="docutils literal notranslate"><span class="pre">get_feature_names</span></code>. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13894">#13894</a>
by <a class="reference external" href="https://github.com/jamesmyatt">James Myatt</a>.</p></li>
</ul>
</section>
<section id="sklearn-utils-sparsefuncs">
<h3><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.utils.sparsefuncs</span></code><a class="headerlink" href="#sklearn-utils-sparsefuncs" title="Permalink to this heading">¶</a></h3>
<ul class="simple">
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> Fixed a bug where <code class="xref py py-func docutils literal notranslate"><span class="pre">min_max_axis</span></code> would fail on 32-bit systems
for certain large inputs. This affects <a class="reference internal" href="../modules/generated/sklearn.preprocessing.MaxAbsScaler.html#sklearn.preprocessing.MaxAbsScaler" title="sklearn.preprocessing.MaxAbsScaler"><code class="xref py py-class docutils literal notranslate"><span class="pre">preprocessing.MaxAbsScaler</span></code></a>,
<a class="reference internal" href="../modules/generated/sklearn.preprocessing.normalize.html#sklearn.preprocessing.normalize" title="sklearn.preprocessing.normalize"><code class="xref py py-func docutils literal notranslate"><span class="pre">preprocessing.normalize</span></code></a> and <a class="reference internal" href="../modules/generated/sklearn.preprocessing.LabelBinarizer.html#sklearn.preprocessing.LabelBinarizer" title="sklearn.preprocessing.LabelBinarizer"><code class="xref py py-class docutils literal notranslate"><span class="pre">preprocessing.LabelBinarizer</span></code></a>.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13741">#13741</a> by <a class="reference external" href="https://github.com/rlms">Roddy MacSween</a>.</p></li>
</ul>
</section>
</section>
</section>
<section id="version-0-21-1">
<span id="changes-0-21-1"></span><h1>Version 0.21.1<a class="headerlink" href="#version-0-21-1" title="Permalink to this heading">¶</a></h1>
<p><strong>17 May 2019</strong></p>
<p>This is a bug-fix release to primarily resolve some packaging issues in version
0.21.0. It also includes minor documentation improvements and some bug fixes.</p>
<section id="id2">
<h2>Changelog<a class="headerlink" href="#id2" title="Permalink to this heading">¶</a></h2>
<section id="id3">
<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="#id3" title="Permalink to this heading">¶</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.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> to only check
classifier and not regressor for the multiclass-multioutput case.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/14309">#14309</a> by <a class="reference external" href="https://github.com/glemaitre">Guillaume Lemaitre</a>.</p></li>
</ul>
</section>
<section id="id4">
<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="#id4" title="Permalink to this heading">¶</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.pairwise_distances.html#sklearn.metrics.pairwise_distances" title="sklearn.metrics.pairwise_distances"><code class="xref py py-class docutils literal notranslate"><span class="pre">metrics.pairwise_distances</span></code></a> where it would raise
<code class="docutils literal notranslate"><span class="pre">AttributeError</span></code> for boolean metrics when <code class="docutils literal notranslate"><span class="pre">X</span></code> had a boolean dtype and
<code class="docutils literal notranslate"><span class="pre">Y</span> <span class="pre">==</span> <span class="pre">None</span></code>.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/issues/13864">#13864</a> by <a class="reference external" href="https://github.com/rick2047">Paresh Mathur</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> Fixed two bugs in <a class="reference internal" href="../modules/generated/sklearn.metrics.pairwise_distances.html#sklearn.metrics.pairwise_distances" title="sklearn.metrics.pairwise_distances"><code class="xref py py-class docutils literal notranslate"><span class="pre">metrics.pairwise_distances</span></code></a> when
<code class="docutils literal notranslate"><span class="pre">n_jobs</span> <span class="pre">></span> <span class="pre">1</span></code>. First it used to return a distance matrix with same dtype as
input, even for integer dtype. Then the diagonal was not zeros for euclidean
metric when <code class="docutils literal notranslate"><span class="pre">Y</span></code> is <code class="docutils literal notranslate"><span class="pre">X</span></code>. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/issues/13877">#13877</a> by
<a class="reference external" href="https://github.com/jeremiedbb">Jérémie du Boisberranger</a>.</p></li>
</ul>
</section>
<section id="id5">
<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="#id5" title="Permalink to this heading">¶</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.neighbors.KernelDensity.html#sklearn.neighbors.KernelDensity" title="sklearn.neighbors.KernelDensity"><code class="xref py py-class docutils literal notranslate"><span class="pre">neighbors.KernelDensity</span></code></a> which could not be
restored from a pickle if <code class="docutils literal notranslate"><span class="pre">sample_weight</span></code> had been used.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/issues/13772">#13772</a> by <a class="reference external" href="https://github.com/aditya1702">Aditya Vyas</a>.</p></li>
</ul>
</section>
</section>
</section>
<section id="version-0-21-0">
<span id="changes-0-21"></span><h1>Version 0.21.0<a class="headerlink" href="#version-0-21-0" title="Permalink to this heading">¶</a></h1>
<p><strong>May 2019</strong></p>
<section id="id6">
<h2>Changed models<a class="headerlink" href="#id6" title="Permalink to this heading">¶</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.discriminant_analysis.LinearDiscriminantAnalysis.html#sklearn.discriminant_analysis.LinearDiscriminantAnalysis" title="sklearn.discriminant_analysis.LinearDiscriminantAnalysis"><code class="xref py py-class docutils literal notranslate"><span class="pre">discriminant_analysis.LinearDiscriminantAnalysis</span></code></a> for multiclass
classification. <span class="raw-html"><span class="badge badge-danger">Fix</span></span> </p></li>
<li><p><a class="reference internal" href="../modules/generated/sklearn.discriminant_analysis.LinearDiscriminantAnalysis.html#sklearn.discriminant_analysis.LinearDiscriminantAnalysis" title="sklearn.discriminant_analysis.LinearDiscriminantAnalysis"><code class="xref py py-class docutils literal notranslate"><span class="pre">discriminant_analysis.LinearDiscriminantAnalysis</span></code></a> with ‘eigen’
solver. <span class="raw-html"><span class="badge badge-danger">Fix</span></span> </p></li>
<li><p><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> <span class="raw-html"><span class="badge badge-danger">Fix</span></span> </p></li>
<li><p>Decision trees and derived ensembles when both <code class="docutils literal notranslate"><span class="pre">max_depth</span></code> and
<code class="docutils literal notranslate"><span class="pre">max_leaf_nodes</span></code> are set. <span class="raw-html"><span class="badge badge-danger">Fix</span></span> </p></li>
<li><p><a class="reference internal" href="../modules/generated/sklearn.linear_model.LogisticRegression.html#sklearn.linear_model.LogisticRegression" title="sklearn.linear_model.LogisticRegression"><code class="xref py py-class docutils literal notranslate"><span class="pre">linear_model.LogisticRegression</span></code></a> and
<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> with ‘saga’ solver. <span class="raw-html"><span class="badge badge-danger">Fix</span></span> </p></li>
<li><p><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> <span class="raw-html"><span class="badge badge-danger">Fix</span></span> </p></li>
<li><p><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">sklearn.feature_extraction.text.HashingVectorizer</span></code></a>,
<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">sklearn.feature_extraction.text.TfidfVectorizer</span></code></a>, and
<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">sklearn.feature_extraction.text.CountVectorizer</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.neural_network.MLPClassifier.html#sklearn.neural_network.MLPClassifier" title="sklearn.neural_network.MLPClassifier"><code class="xref py py-class docutils literal notranslate"><span class="pre">neural_network.MLPClassifier</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.svm.SVC.html#sklearn.svm.SVC.decision_function" title="sklearn.svm.SVC.decision_function"><code class="xref py py-func docutils literal notranslate"><span class="pre">svm.SVC.decision_function</span></code></a> and
<a class="reference internal" href="../modules/generated/sklearn.multiclass.OneVsOneClassifier.html#sklearn.multiclass.OneVsOneClassifier.decision_function" title="sklearn.multiclass.OneVsOneClassifier.decision_function"><code class="xref py py-func docutils literal notranslate"><span class="pre">multiclass.OneVsOneClassifier.decision_function</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.linear_model.SGDClassifier.html#sklearn.linear_model.SGDClassifier" title="sklearn.linear_model.SGDClassifier"><code class="xref py py-class docutils literal notranslate"><span class="pre">linear_model.SGDClassifier</span></code></a> and any derived classifiers. <span class="raw-html"><span class="badge badge-danger">Fix</span></span> </p></li>
<li><p>Any model using the <code class="xref py py-func docutils literal notranslate"><span class="pre">linear_model._sag.sag_solver</span></code> function with a <code class="docutils literal notranslate"><span class="pre">0</span></code>
seed, including <a class="reference internal" href="../modules/generated/sklearn.linear_model.LogisticRegression.html#sklearn.linear_model.LogisticRegression" title="sklearn.linear_model.LogisticRegression"><code class="xref py py-class docutils literal notranslate"><span class="pre">linear_model.LogisticRegression</span></code></a>,
<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>, <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>,
and <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> with ‘sag’ solver. <span class="raw-html"><span class="badge badge-danger">Fix</span></span> </p></li>
<li><p><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> when using leave-one-out cross-validation
with sparse inputs. <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>
</section>
<section id="known-major-bugs">
<h2>Known Major Bugs<a class="headerlink" href="#known-major-bugs" title="Permalink to this heading">¶</a></h2>
<ul class="simple">
<li><p>The default <code class="docutils literal notranslate"><span class="pre">max_iter</span></code> for <a class="reference internal" href="../modules/generated/sklearn.linear_model.LogisticRegression.html#sklearn.linear_model.LogisticRegression" title="sklearn.linear_model.LogisticRegression"><code class="xref py py-class docutils literal notranslate"><span class="pre">linear_model.LogisticRegression</span></code></a> is too
small for many solvers given the default <code class="docutils literal notranslate"><span class="pre">tol</span></code>. In particular, we
accidentally changed the default <code class="docutils literal notranslate"><span class="pre">max_iter</span></code> for the liblinear solver from
1000 to 100 iterations in <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/3591">#3591</a> released in version 0.16.
In a future release we hope to choose better default <code class="docutils literal notranslate"><span class="pre">max_iter</span></code> and <code class="docutils literal notranslate"><span class="pre">tol</span></code>
heuristically depending on the solver (see <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13317">#13317</a>).</p></li>
</ul>
</section>
<section id="id7">
<h2>Changelog<a class="headerlink" href="#id7" title="Permalink to this heading">¶</a></h2>
<p>Support for Python 3.4 and below has been officially dropped.</p>
<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 heading">¶</a></h3>
<ul class="simple">
<li><p><span class="raw-html"><span class="badge badge-warning">API Change</span></span> The R2 score used when calling <code class="docutils literal notranslate"><span class="pre">score</span></code> on a regressor will use
<code class="docutils literal notranslate"><span class="pre">multioutput='uniform_average'</span></code> from version 0.23 to keep consistent with
<a class="reference internal" href="../modules/generated/sklearn.metrics.r2_score.html#sklearn.metrics.r2_score" title="sklearn.metrics.r2_score"><code class="xref py py-func docutils literal notranslate"><span class="pre">metrics.r2_score</span></code></a>. This will influence the <code class="docutils literal notranslate"><span class="pre">score</span></code> method of all
the multioutput regressors (except for
<a class="reference internal" href="../modules/generated/sklearn.multioutput.MultiOutputRegressor.html#sklearn.multioutput.MultiOutputRegressor" title="sklearn.multioutput.MultiOutputRegressor"><code class="xref py py-class docutils literal notranslate"><span class="pre">multioutput.MultiOutputRegressor</span></code></a>).
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13157">#13157</a> by <a class="reference external" href="https://github.com/qinhanmin2014">Hanmin Qin</a>.</p></li>
</ul>
</section>
<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 heading">¶</a></h3>
<ul class="simple">
<li><p><span class="raw-html"><span class="badge badge-info">Enhancement</span></span> Added support to bin the data passed into
<a class="reference internal" href="../modules/generated/sklearn.calibration.calibration_curve.html#sklearn.calibration.calibration_curve" title="sklearn.calibration.calibration_curve"><code class="xref py py-class docutils literal notranslate"><span class="pre">calibration.calibration_curve</span></code></a> by quantiles instead of uniformly
between 0 and 1.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13086">#13086</a> by <a class="reference external" href="https://github.com/srcole">Scott Cole</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-info">Enhancement</span></span> Allow n-dimensional arrays as input for
<code class="docutils literal notranslate"><span class="pre">calibration.CalibratedClassifierCV</span></code>. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13485">#13485</a> by
<a class="reference external" href="https://github.com/wdevazelhes">William de Vazelhes</a>.</p></li>
</ul>
</section>
<section id="id8">
<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="#id8" title="Permalink to this heading">¶</a></h3>
<ul class="simple">
<li><p><span class="raw-html"><span class="badge badge-success">Major Feature</span></span> A new clustering algorithm: <a class="reference internal" href="../modules/generated/sklearn.cluster.OPTICS.html#sklearn.cluster.OPTICS" title="sklearn.cluster.OPTICS"><code class="xref py py-class docutils literal notranslate"><span class="pre">cluster.OPTICS</span></code></a>: an
algorithm related to <a class="reference internal" href="../modules/generated/sklearn.cluster.DBSCAN.html#sklearn.cluster.DBSCAN" title="sklearn.cluster.DBSCAN"><code class="xref py py-class docutils literal notranslate"><span class="pre">cluster.DBSCAN</span></code></a>, that has hyperparameters easier
to set and that scales better, by <a class="reference external" href="https://github.com/espg">Shane</a>,
<a class="reference external" href="https://github.com/adrinjalali">Adrin Jalali</a>, <a class="reference external" href="https://github.com/kno10">Erich Schubert</a>, <a class="reference external" href="https://github.com/qinhanmin2014">Hanmin Qin</a>, and
<a class="reference external" href="https://github.com/assiaben">Assia Benbihi</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.Birch.html#sklearn.cluster.Birch" title="sklearn.cluster.Birch"><code class="xref py py-class docutils literal notranslate"><span class="pre">cluster.Birch</span></code></a> could occasionally raise an
AttributeError. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13651">#13651</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> Fixed a bug 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> where empty clusters weren’t
correctly relocated when using sample weights. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13486">#13486</a> by
<a class="reference external" href="https://github.com/jeremiedbb">Jérémie du Boisberranger</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">n_components_</span></code> attribute in <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> has been renamed to
<code class="docutils literal notranslate"><span class="pre">n_connected_components_</span></code>.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13427">#13427</a> by <a class="reference external" href="https://github.com/scouvreur">Stephane Couvreur</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.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 accept a <code class="docutils literal notranslate"><span class="pre">distance_threshold</span></code>
parameter which can be used to find the clusters instead of <code class="docutils literal notranslate"><span class="pre">n_clusters</span></code>.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/issues/9069">#9069</a> by <a class="reference external" href="https://github.com/VathsalaAchar">Vathsala Achar</a> and <a class="reference external" href="https://github.com/adrinjalali">Adrin Jalali</a>.</p></li>
</ul>
</section>
<section id="id9">
<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="#id9" title="Permalink to this heading">¶</a></h3>
<ul class="simple">
<li><p><span class="raw-html"><span class="badge badge-warning">API Change</span></span> <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> is no longer an experimental
feature. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13835">#13835</a> by <a class="reference external" href="https://github.com/qinhanmin2014">Hanmin Qin</a>.</p></li>
</ul>
</section>
<section id="id10">
<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="#id10" title="Permalink to this heading">¶</a></h3>
<ul class="simple">
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> Added support for 64-bit group IDs and pointers in SVMLight files.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/10727">#10727</a> by <a class="reference external" href="https://github.com/bryan-woods">Bryan K Woods</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.datasets.load_sample_images.html#sklearn.datasets.load_sample_images" title="sklearn.datasets.load_sample_images"><code class="xref py py-func docutils literal notranslate"><span class="pre">datasets.load_sample_images</span></code></a> returns images with a deterministic
order. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13250">#13250</a> by <a class="reference external" href="https://github.com/thomasjpfan">Thomas Fan</a>.</p></li>
</ul>
</section>
<section id="id11">
<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="#id11" title="Permalink to this heading">¶</a></h3>
<ul class="simple">
<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 has deterministic output
(resolved sign ambiguity in eigenvalue decomposition of the kernel matrix).
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13241">#13241</a> by <a class="reference external" href="https://github.com/bellet">Aurélien Bellet</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.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>, <code class="docutils literal notranslate"><span class="pre">fit().transform()</span></code>
now produces the correct output (the same as <code class="docutils literal notranslate"><span class="pre">fit_transform()</span></code>) in case
of non-removed zero eigenvalues (<code class="docutils literal notranslate"><span class="pre">remove_zero_eig=False</span></code>).
<code class="docutils literal notranslate"><span class="pre">fit_inverse_transform</span></code> was also accelerated by using the same trick as
<code class="docutils literal notranslate"><span class="pre">fit_transform</span></code> to compute the transform of <code class="docutils literal notranslate"><span class="pre">X</span></code>.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/12143">#12143</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 in <a class="reference internal" href="../modules/generated/sklearn.decomposition.NMF.html#sklearn.decomposition.NMF" title="sklearn.decomposition.NMF"><code class="xref py py-class docutils literal notranslate"><span class="pre">decomposition.NMF</span></code></a> where <code class="docutils literal notranslate"><span class="pre">init</span> <span class="pre">=</span> <span class="pre">'nndsvd'</span></code>,
<code class="docutils literal notranslate"><span class="pre">init</span> <span class="pre">=</span> <span class="pre">'nndsvda'</span></code>, and <code class="docutils literal notranslate"><span class="pre">init</span> <span class="pre">=</span> <span class="pre">'nndsvdar'</span></code> are allowed when
<code class="docutils literal notranslate"><span class="pre">n_components</span> <span class="pre"><</span> <span class="pre">n_features</span></code> instead of
<code class="docutils literal notranslate"><span class="pre">n_components</span> <span class="pre"><=</span> <span class="pre">min(n_samples,</span> <span class="pre">n_features)</span></code>.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/11650">#11650</a> by <a class="reference external" href="https://github.com/hossein-pourbozorg">Hossein Pourbozorg</a> and
<a class="reference external" href="https://github.com/zjpoh">Zijie (ZJ) Poh</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="code docutils literal notranslate"><span class="pre">init</span></code> argument in
<a class="reference internal" href="../modules/generated/sklearn.decomposition.non_negative_factorization.html#sklearn.decomposition.non_negative_factorization" title="sklearn.decomposition.non_negative_factorization"><code class="xref py py-func docutils literal notranslate"><span class="pre">decomposition.non_negative_factorization</span></code></a> will change from
<code class="code docutils literal notranslate"><span class="pre">random</span></code> to <code class="code docutils literal notranslate"><span class="pre">None</span></code> in version 0.23 to make it consistent with
<a class="reference internal" href="../modules/generated/sklearn.decomposition.NMF.html#sklearn.decomposition.NMF" title="sklearn.decomposition.NMF"><code class="xref py py-class docutils literal notranslate"><span class="pre">decomposition.NMF</span></code></a>. A FutureWarning is raised when
the default value is used.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/12988">#12988</a> by <a class="reference external" href="https://github.com/zjpoh">Zijie (ZJ) Poh</a>.</p></li>
</ul>
</section>
<section id="sklearn-discriminant-analysis">
<h3><a class="reference internal" href="../modules/classes.html#module-sklearn.discriminant_analysis" title="sklearn.discriminant_analysis"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.discriminant_analysis</span></code></a><a class="headerlink" href="#sklearn-discriminant-analysis" title="Permalink to this heading">¶</a></h3>
<ul class="simple">
<li><p><span class="raw-html"><span class="badge badge-info">Enhancement</span></span> <a class="reference internal" href="../modules/generated/sklearn.discriminant_analysis.LinearDiscriminantAnalysis.html#sklearn.discriminant_analysis.LinearDiscriminantAnalysis" title="sklearn.discriminant_analysis.LinearDiscriminantAnalysis"><code class="xref py py-class docutils literal notranslate"><span class="pre">discriminant_analysis.LinearDiscriminantAnalysis</span></code></a> now
preserves <code class="docutils literal notranslate"><span class="pre">float32</span></code> and <code class="docutils literal notranslate"><span class="pre">float64</span></code> dtypes. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/8769">#8769</a> and
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/11000">#11000</a> by <a class="reference external" href="https://github.com/thibsej">Thibault Sejourne</a></p></li>
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> A <code class="docutils literal notranslate"><span class="pre">ChangedBehaviourWarning</span></code> is now raised when
<a class="reference internal" href="../modules/generated/sklearn.discriminant_analysis.LinearDiscriminantAnalysis.html#sklearn.discriminant_analysis.LinearDiscriminantAnalysis" title="sklearn.discriminant_analysis.LinearDiscriminantAnalysis"><code class="xref py py-class docutils literal notranslate"><span class="pre">discriminant_analysis.LinearDiscriminantAnalysis</span></code></a> is given as
parameter <code class="docutils literal notranslate"><span class="pre">n_components</span> <span class="pre">></span> <span class="pre">min(n_features,</span> <span class="pre">n_classes</span> <span class="pre">-</span> <span class="pre">1)</span></code>, and
<code class="docutils literal notranslate"><span class="pre">n_components</span></code> is changed to <code class="docutils literal notranslate"><span class="pre">min(n_features,</span> <span class="pre">n_classes</span> <span class="pre">-</span> <span class="pre">1)</span></code> if so.
Previously the change was made, but silently. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/11526">#11526</a> by
<a class="reference external" href="https://github.com/wdevazelhes">William de Vazelhes</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.discriminant_analysis.LinearDiscriminantAnalysis.html#sklearn.discriminant_analysis.LinearDiscriminantAnalysis" title="sklearn.discriminant_analysis.LinearDiscriminantAnalysis"><code class="xref py py-class docutils literal notranslate"><span class="pre">discriminant_analysis.LinearDiscriminantAnalysis</span></code></a>
where the predicted probabilities would be incorrectly computed in the
multiclass case. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/6848">#6848</a>, by <a class="reference external" href="https://github.com/agamemnonc">Agamemnon Krasoulis</a> and <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> Fixed a bug in <a class="reference internal" href="../modules/generated/sklearn.discriminant_analysis.LinearDiscriminantAnalysis.html#sklearn.discriminant_analysis.LinearDiscriminantAnalysis" title="sklearn.discriminant_analysis.LinearDiscriminantAnalysis"><code class="xref py py-class docutils literal notranslate"><span class="pre">discriminant_analysis.LinearDiscriminantAnalysis</span></code></a>
where the predicted probabilities would be incorrectly computed with <code class="docutils literal notranslate"><span class="pre">eigen</span></code>
solver. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/11727">#11727</a>, by <a class="reference external" href="https://github.com/agamemnonc">Agamemnon Krasoulis</a>.</p></li>
</ul>
</section>
<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 heading">¶</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.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> where the
<code class="docutils literal notranslate"><span class="pre">predict_proba</span></code> method was returning int32 array instead of
float64 for the <code class="docutils literal notranslate"><span class="pre">stratified</span></code> strategy. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13266">#13266</a> by
<a class="reference external" href="https://github.com/chkoar">Christos Aridas</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.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> where it was throwing a
dimension mismatch error in prediction time if a column vector <code class="docutils literal notranslate"><span class="pre">y</span></code> with
<code class="docutils literal notranslate"><span class="pre">shape=(n,</span> <span class="pre">1)</span></code> was given at <code class="docutils literal notranslate"><span class="pre">fit</span></code> time. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13545">#13545</a> by <a class="reference external" href="https://github.com/nsorros">Nick
Sorros</a> and <a class="reference external" href="https://github.com/adrinjalali">Adrin Jalali</a>.</p></li>
</ul>
</section>
<section id="id12">
<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="#id12" title="Permalink to this heading">¶</a></h3>
<ul>
<li><p><span class="raw-html"><span class="badge badge-success">Major Feature</span></span> Add two new implementations of
gradient boosting trees: <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>. The implementation of
these estimators is inspired by
<a class="reference external" href="https://github.com/Microsoft/LightGBM">LightGBM</a> and can be orders of
magnitude faster than <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
<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> when the number of samples is
larger than tens of thousands of samples. The API of these new estimators
is slightly different, and some of the features from
<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> are not yet supported.</p>
<p>These new estimators are experimental, which means that their results or
their API might change without any deprecation cycle. To use them, you
need to explicitly import <code class="docutils literal notranslate"><span class="pre">enable_hist_gradient_boosting</span></code>:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="c1"># explicitly require this experimental feature</span>
<span class="gp">>>> </span><span class="kn">from</span> <span class="nn">sklearn.experimental</span> <span class="kn">import</span> <span class="n">enable_hist_gradient_boosting</span> <span class="c1"># noqa</span>
<span class="gp">>>> </span><span class="c1"># now you can import normally from sklearn.ensemble</span>
<span class="gp">>>> </span><span class="kn">from</span> <span class="nn">sklearn.ensemble</span> <span class="kn">import</span> <span class="n">HistGradientBoostingClassifier</span>
</pre></div>
</div>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>Update: since version 1.0, these estimators are not experimental
anymore and you don’t need to use <code class="docutils literal notranslate"><span class="pre">from</span> <span class="pre">sklearn.experimental</span> <span class="pre">import</span>
<span class="pre">enable_hist_gradient_boosting</span></code>.</p>
</div>
<p><a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/12807">#12807</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-success">Feature</span></span> Add <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>
which provides an equivalent of <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>
for regression problems.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/12513">#12513</a> by <a class="reference external" href="https://github.com/stsouko">Ramil Nugmanov</a> and
<a class="reference external" href="https://github.com/mohamed-ali">Mohamed Ali Jamaoui</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-info">Efficiency</span></span> Make <a class="reference internal" href="../modules/generated/sklearn.ensemble.IsolationForest.html#sklearn.ensemble.IsolationForest" title="sklearn.ensemble.IsolationForest"><code class="xref py py-class docutils literal notranslate"><span class="pre">ensemble.IsolationForest</span></code></a> prefer threads over
processes when running with <code class="docutils literal notranslate"><span class="pre">n_jobs</span> <span class="pre">></span> <span class="pre">1</span></code> as the underlying decision tree
fit calls do release the GIL. This changes reduces memory usage and
communication overhead. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/12543">#12543</a> by <a class="reference external" href="https://github.com/istorch">Isaac Storch</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-info">Efficiency</span></span> Make <a class="reference internal" href="../modules/generated/sklearn.ensemble.IsolationForest.html#sklearn.ensemble.IsolationForest" title="sklearn.ensemble.IsolationForest"><code class="xref py py-class docutils literal notranslate"><span class="pre">ensemble.IsolationForest</span></code></a> more memory efficient
by avoiding keeping in memory each tree prediction. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13260">#13260</a> by
<a class="reference external" href="https://ngoix.github.io/">Nicolas Goix</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.ensemble.IsolationForest.html#sklearn.ensemble.IsolationForest" title="sklearn.ensemble.IsolationForest"><code class="xref py py-class docutils literal notranslate"><span class="pre">ensemble.IsolationForest</span></code></a> now uses chunks of data at
prediction step, thus capping the memory usage. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13283">#13283</a> by
<a class="reference external" href="https://ngoix.github.io/">Nicolas Goix</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.ensemble.GradientBoostingClassifier.html#sklearn.ensemble.GradientBoostingClassifier" title="sklearn.ensemble.GradientBoostingClassifier"><code class="xref py py-class docutils literal notranslate"><span class="pre">sklearn.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">sklearn.ensemble.GradientBoostingRegressor</span></code></a> now keep the
input <code class="docutils literal notranslate"><span class="pre">y</span></code> as <code class="docutils literal notranslate"><span class="pre">float64</span></code> to avoid it being copied internally by trees.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13524">#13524</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> Minimized the validation of X in
<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> and <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>
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13174">#13174</a> by <a class="reference external" href="https://github.com/chkoar">Christos Aridas</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.IsolationForest.html#sklearn.ensemble.IsolationForest" title="sklearn.ensemble.IsolationForest"><code class="xref py py-class docutils literal notranslate"><span class="pre">ensemble.IsolationForest</span></code></a> now exposes <code class="docutils literal notranslate"><span class="pre">warm_start</span></code>
parameter, allowing iterative addition of trees to an isolation
forest. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13496">#13496</a> by <a class="reference external" href="https://github.com/petibear">Peter Marko</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> The values of <code class="docutils literal notranslate"><span class="pre">feature_importances_</span></code> in all random forest based
models (i.e.
<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 internal" href="../modules/generated/sklearn.ensemble.RandomTreesEmbedding.html#sklearn.ensemble.RandomTreesEmbedding" title="sklearn.ensemble.RandomTreesEmbedding"><code class="xref py py-class docutils literal notranslate"><span class="pre">ensemble.RandomTreesEmbedding</span></code></a>,
<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>) now:</p>
<ul class="simple">
<li><p>sum up to <code class="docutils literal notranslate"><span class="pre">1</span></code></p></li>
<li><p>all the single node trees in feature importance calculation are ignored</p></li>
<li><p>in case all trees have only one single node (i.e. a root node),
feature importances will be an array of all zeros.</p></li>
</ul>
<p><a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13636">#13636</a> and <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13620">#13620</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> Fixed a bug 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>, which didn’t support
scikit-learn estimators as the initial estimator. Also added support of
initial estimator which does not support sample weights. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/12436">#12436</a> by
<a class="reference external" href="https://github.com/jeremiedbb">Jérémie du Boisberranger</a> and <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/12983">#12983</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 the output of the average path length computed in
<a class="reference internal" href="../modules/generated/sklearn.ensemble.IsolationForest.html#sklearn.ensemble.IsolationForest" title="sklearn.ensemble.IsolationForest"><code class="xref py py-class docutils literal notranslate"><span class="pre">ensemble.IsolationForest</span></code></a> when the input is either 0, 1 or 2.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13251">#13251</a> by <a class="reference external" href="https://github.com/albertcthomas">Albert Thomas</a>
and <a class="reference external" href="https://github.com/joshuakennethjones">joshuakennethjones</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.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> where
the gradients would be incorrectly computed in multiclass classification
problems. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/12715">#12715</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 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> where
validation sets for early stopping were not sampled with stratification.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13164">#13164</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 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> where
the default initial prediction of a multiclass classifier would predict the
classes priors instead of the log of the priors. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/12983">#12983</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 in <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> where the
<code class="docutils literal notranslate"><span class="pre">predict</span></code> method would error for multiclass multioutput forests models
if any targets were strings. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/12834">#12834</a> by <a class="reference external" href="https://github.com/elsander">Elizabeth Sander</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> Fixed a bug in <code class="xref py py-class docutils literal notranslate"><span class="pre">ensemble.gradient_boosting.LossFunction</span></code> and
<code class="xref py py-class docutils literal notranslate"><span class="pre">ensemble.gradient_boosting.LeastSquaresError</span></code> where the default
value of <code class="docutils literal notranslate"><span class="pre">learning_rate</span></code> in <code class="docutils literal notranslate"><span class="pre">update_terminal_regions</span></code> is not consistent
with the document and the caller functions. Note however that directly using
these loss functions is deprecated.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/6463">#6463</a> by <a class="reference external" href="https://github.com/movelikeriver">movelikeriver</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">ensemble.partial_dependence</span></code> (and consequently the new
version <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">sklearn.inspection.partial_dependence</span></code></a>) now takes sample
weights into account for the partial dependence computation when the
gradient boosting model has been trained with sample weights.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13193">#13193</a> by <a class="reference external" href="https://github.com/samronsin">Samuel O. Ronsin</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-warning">API Change</span></span> <code class="xref py py-func docutils literal notranslate"><span class="pre">ensemble.partial_dependence</span></code> and
<code class="xref py py-func docutils literal notranslate"><span class="pre">ensemble.plot_partial_dependence</span></code> are now deprecated in favor of
<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
<code class="xref py py-func docutils literal notranslate"><span class="pre">inspection.plot_partial_dependence</span></code>.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/12599">#12599</a> by <a class="reference external" href="https://github.com/trevorstephens">Trevor Stephens</a> and
<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.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> were failing during <code class="docutils literal notranslate"><span class="pre">fit</span></code> in one
of the estimators was set to <code class="docutils literal notranslate"><span class="pre">None</span></code> and <code class="docutils literal notranslate"><span class="pre">sample_weight</span></code> was not <code class="docutils literal notranslate"><span class="pre">None</span></code>.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13779">#13779</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> <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> accept <code class="docutils literal notranslate"><span class="pre">'drop'</span></code> to disable an estimator
in addition to <code class="docutils literal notranslate"><span class="pre">None</span></code> to be consistent with other estimators (i.e.,
<a class="reference internal" href="../modules/generated/sklearn.pipeline.FeatureUnion.html#sklearn.pipeline.FeatureUnion" title="sklearn.pipeline.FeatureUnion"><code class="xref py py-class docutils literal notranslate"><span class="pre">pipeline.FeatureUnion</span></code></a> and <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>).
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13780">#13780</a> by <a class="reference external" href="https://github.com/glemaitre">Guillaume Lemaitre</a>.</p></li>
</ul>
</section>
<section id="sklearn-externals">
<h3><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.externals</span></code><a class="headerlink" href="#sklearn-externals" title="Permalink to this heading">¶</a></h3>
<ul class="simple">
<li><p><span class="raw-html"><span class="badge badge-warning">API Change</span></span> Deprecated <code class="xref py py-mod docutils literal notranslate"><span class="pre">externals.six</span></code> since we have dropped support for
Python 2.7. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/12916">#12916</a> by <a class="reference external" href="https://github.com/qinhanmin2014">Hanmin Qin</a>.</p></li>
</ul>
</section>
<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 heading">¶</a></h3>
<ul class="simple">
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> If <code class="docutils literal notranslate"><span class="pre">input='file'</span></code> or <code class="docutils literal notranslate"><span class="pre">input='filename'</span></code>, and a callable is given as
the <code class="docutils literal notranslate"><span class="pre">analyzer</span></code>, <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">sklearn.feature_extraction.text.HashingVectorizer</span></code></a>,
<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">sklearn.feature_extraction.text.TfidfVectorizer</span></code></a>, and
<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">sklearn.feature_extraction.text.CountVectorizer</span></code></a> now read the data
from the file(s) and then pass it to the given <code class="docutils literal notranslate"><span class="pre">analyzer</span></code>, instead of
passing the file name(s) or the file object(s) to the analyzer.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13641">#13641</a> by <a class="reference external" href="https://github.com/adrinjalali">Adrin Jalali</a>.</p></li>
</ul>
</section>
<section id="id13">
<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="#id13" title="Permalink to this heading">¶</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.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>, which is a strategy
for imputing missing values by modeling each feature with missing values as a
function of other features in a round-robin fashion. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/8478">#8478</a> and
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/12177">#12177</a> by <a class="reference external" href="https://github.com/sergeyf">Sergey Feldman</a> and <a class="reference external" href="https://github.com/benlawson">Ben Lawson</a>.</p>
<p>The API of IterativeImputer is experimental and subject to change without any
deprecation cycle. To use them, you need to explicitly import
<code class="docutils literal notranslate"><span class="pre">enable_iterative_imputer</span></code>:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="kn">from</span> <span class="nn">sklearn.experimental</span> <span class="kn">import</span> <span class="n">enable_iterative_imputer</span> <span class="c1"># noqa</span>
<span class="gp">>>> </span><span class="c1"># now you can import normally from sklearn.impute</span>
<span class="gp">>>> </span><span class="kn">from</span> <span class="nn">sklearn.impute</span> <span class="kn">import</span> <span class="n">IterativeImputer</span>
</pre></div>
</div>
</li>
<li><p><span class="raw-html"><span class="badge badge-success">Feature</span></span> The <a class="reference internal" href="../modules/generated/sklearn.impute.SimpleImputer.html#sklearn.impute.SimpleImputer" title="sklearn.impute.SimpleImputer"><code class="xref py py-class docutils literal notranslate"><span class="pre">impute.SimpleImputer</span></code></a> and
<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> have a new parameter <code class="docutils literal notranslate"><span class="pre">'add_indicator'</span></code>,
which simply stacks a <a class="reference internal" href="../modules/generated/sklearn.impute.MissingIndicator.html#sklearn.impute.MissingIndicator" title="sklearn.impute.MissingIndicator"><code class="xref py py-class docutils literal notranslate"><span class="pre">impute.MissingIndicator</span></code></a> transform into the
output of the imputer’s transform. That allows a predictive estimator to
account for missingness. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/12583">#12583</a>, <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13601">#13601</a> by <a class="reference external" href="https://github.com/DanilBaibak">Danylo Baibak</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> In <a class="reference internal" href="../modules/generated/sklearn.impute.MissingIndicator.html#sklearn.impute.MissingIndicator" title="sklearn.impute.MissingIndicator"><code class="xref py py-class docutils literal notranslate"><span class="pre">impute.MissingIndicator</span></code></a> avoid implicit densification by
raising an exception if input is sparse add <code class="docutils literal notranslate"><span class="pre">missing_values</span></code> property
is set to 0. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13240">#13240</a> by <a class="reference external" href="https://github.com/btel">Bartosz Telenczuk</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> Fixed two bugs in <a class="reference internal" href="../modules/generated/sklearn.impute.MissingIndicator.html#sklearn.impute.MissingIndicator" title="sklearn.impute.MissingIndicator"><code class="xref py py-class docutils literal notranslate"><span class="pre">impute.MissingIndicator</span></code></a>. First, when
<code class="docutils literal notranslate"><span class="pre">X</span></code> is sparse, all the non-zero non missing values used to become
explicit False in the transformed data. Then, when
<code class="docutils literal notranslate"><span class="pre">features='missing-only'</span></code>, all features used to be kept if there were no
missing values at all. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13562">#13562</a> by <a class="reference external" href="https://github.com/jeremiedbb">Jérémie du Boisberranger</a>.</p></li>
</ul>
</section>
<section id="id14">
<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="#id14" title="Permalink to this heading">¶</a></h3>
<p>(new subpackage)</p>
<ul class="simple">
<li><p><span class="raw-html"><span class="badge badge-success">Feature</span></span> Partial dependence plots
(<code class="xref py py-func docutils literal notranslate"><span class="pre">inspection.plot_partial_dependence</span></code>) are now supported for
any regressor or classifier (provided that they have a <code class="docutils literal notranslate"><span class="pre">predict_proba</span></code>
method). <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/12599">#12599</a> by <a class="reference external" href="https://github.com/trevorstephens">Trevor Stephens</a> and
<a class="reference external" href="https://github.com/NicolasHug">Nicolas Hug</a>.</p></li>
</ul>
</section>
<section id="sklearn-isotonic">
<h3><a class="reference internal" href="../modules/classes.html#module-sklearn.isotonic" title="sklearn.isotonic"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.isotonic</span></code></a><a class="headerlink" href="#sklearn-isotonic" title="Permalink to this heading">¶</a></h3>
<ul class="simple">
<li><p><span class="raw-html"><span class="badge badge-success">Feature</span></span> Allow different dtypes (such as float32) in
<a class="reference internal" href="../modules/generated/sklearn.isotonic.IsotonicRegression.html#sklearn.isotonic.IsotonicRegression" title="sklearn.isotonic.IsotonicRegression"><code class="xref py py-class docutils literal notranslate"><span class="pre">isotonic.IsotonicRegression</span></code></a>.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/8769">#8769</a> by <a class="reference external" href="https://github.com/vene">Vlad Niculae</a></p></li>
</ul>
</section>
<section id="id15">
<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="#id15" title="Permalink to this heading">¶</a></h3>
<ul class="simple">
<li><p><span class="raw-html"><span class="badge badge-info">Enhancement</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 preserves <code class="docutils literal notranslate"><span class="pre">float32</span></code> and
<code class="docutils literal notranslate"><span class="pre">float64</span></code> dtypes. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/issues/8769">#8769</a> and <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/issues/11000">#11000</a> by
<a class="reference external" href="https://github.com/glemaitre">Guillaume Lemaitre</a>, and <a class="reference external" href="https://github.com/massich">Joan Massich</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.linear_model.LogisticRegression.html#sklearn.linear_model.LogisticRegression" title="sklearn.linear_model.LogisticRegression"><code class="xref py py-class docutils literal notranslate"><span class="pre">linear_model.LogisticRegression</span></code></a> and
<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> now support Elastic-Net penalty,
with the ‘saga’ solver. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/11646">#11646</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-success">Feature</span></span> Added <a class="reference internal" href="../modules/generated/sklearn.linear_model.lars_path_gram.html#sklearn.linear_model.lars_path_gram" title="sklearn.linear_model.lars_path_gram"><code class="xref py py-class docutils literal notranslate"><span class="pre">linear_model.lars_path_gram</span></code></a>, which is
<a class="reference internal" href="../modules/generated/sklearn.linear_model.lars_path.html#sklearn.linear_model.lars_path" title="sklearn.linear_model.lars_path"><code class="xref py py-class docutils literal notranslate"><span class="pre">linear_model.lars_path</span></code></a> in the sufficient stats mode, allowing
users to compute <a class="reference internal" href="../modules/generated/sklearn.linear_model.lars_path.html#sklearn.linear_model.lars_path" title="sklearn.linear_model.lars_path"><code class="xref py py-class docutils literal notranslate"><span class="pre">linear_model.lars_path</span></code></a> without providing
<code class="docutils literal notranslate"><span class="pre">X</span></code> and <code class="docutils literal notranslate"><span class="pre">y</span></code>. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/11699">#11699</a> by <a class="reference external" href="https://github.com/yukuairoy">Kuai Yu</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-info">Efficiency</span></span> <code class="xref py py-func docutils literal notranslate"><span class="pre">linear_model.make_dataset</span></code> now preserves
<code class="docutils literal notranslate"><span class="pre">float32</span></code> and <code class="docutils literal notranslate"><span class="pre">float64</span></code> dtypes, reducing memory consumption in stochastic
gradient, SAG and SAGA solvers.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/8769">#8769</a> and <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/11000">#11000</a> by
<a class="reference external" href="https://github.com/NelleV">Nelle Varoquaux</a>, <a class="reference external" href="https://github.com/Henley13">Arthur Imbert</a>,
<a class="reference external" href="https://github.com/glemaitre">Guillaume Lemaitre</a>, and <a class="reference external" href="https://github.com/massich">Joan Massich</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.LogisticRegression.html#sklearn.linear_model.LogisticRegression" title="sklearn.linear_model.LogisticRegression"><code class="xref py py-class docutils literal notranslate"><span class="pre">linear_model.LogisticRegression</span></code></a> now supports an
unregularized objective when <code class="docutils literal notranslate"><span class="pre">penalty='none'</span></code> is passed. This is
equivalent to setting <code class="docutils literal notranslate"><span class="pre">C=np.inf</span></code> with l2 regularization. Not supported
by the liblinear solver. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/12860">#12860</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> <code class="docutils literal notranslate"><span class="pre">sparse_cg</span></code> solver in <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 supports fitting the intercept (i.e. <code class="docutils literal notranslate"><span class="pre">fit_intercept=True</span></code>) when
inputs are sparse. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13336">#13336</a> by <a class="reference external" href="https://github.com/btel">Bartosz Telenczuk</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-info">Enhancement</span></span> The coordinate descent solver used in <code class="docutils literal notranslate"><span class="pre">Lasso</span></code>, <code class="docutils literal notranslate"><span class="pre">ElasticNet</span></code>,
etc. now issues a <code class="docutils literal notranslate"><span class="pre">ConvergenceWarning</span></code> when it completes without meeting the
desired toleranbce.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/11754">#11754</a> and <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13397">#13397</a> by <a class="reference external" href="https://github.com/brentfagan">Brent Fagan</a> and
<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> Fixed a bug in <a class="reference internal" href="../modules/generated/sklearn.linear_model.LogisticRegression.html#sklearn.linear_model.LogisticRegression" title="sklearn.linear_model.LogisticRegression"><code class="xref py py-class docutils literal notranslate"><span class="pre">linear_model.LogisticRegression</span></code></a> and
<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> with ‘saga’ solver, where the
weights would not be correctly updated in some cases.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/11646">#11646</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-danger">Fix</span></span> Fixed the posterior mean, posterior covariance and returned
regularization parameters in <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>. The
posterior mean and the posterior covariance were not the ones computed
with the last update of the regularization parameters and the returned
regularization parameters were not the final ones. Also fixed the formula of
the log marginal likelihood used to compute the score when
<code class="docutils literal notranslate"><span class="pre">compute_score=True</span></code>. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/12174">#12174</a> by
<a class="reference external" href="https://github.com/albertcthomas">Albert Thomas</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.LassoLarsIC.html#sklearn.linear_model.LassoLarsIC" title="sklearn.linear_model.LassoLarsIC"><code class="xref py py-class docutils literal notranslate"><span class="pre">linear_model.LassoLarsIC</span></code></a>, where user input
<code class="docutils literal notranslate"><span class="pre">copy_X=False</span></code> at instance creation would be overridden by default
parameter value <code class="docutils literal notranslate"><span class="pre">copy_X=True</span></code> in <code class="docutils literal notranslate"><span class="pre">fit</span></code>.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/12972">#12972</a> by <a class="reference external" href="https://github.com/luk-f-a">Lucio Fernandez-Arjona</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.LinearRegression.html#sklearn.linear_model.LinearRegression" title="sklearn.linear_model.LinearRegression"><code class="xref py py-class docutils literal notranslate"><span class="pre">linear_model.LinearRegression</span></code></a> that
was not returning the same coeffecients and intercepts with
<code class="docutils literal notranslate"><span class="pre">fit_intercept=True</span></code> in sparse and dense case.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13279">#13279</a> by <a class="reference external" href="http://alexandre.gramfort.net">Alexandre Gramfort</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.HuberRegressor.html#sklearn.linear_model.HuberRegressor" title="sklearn.linear_model.HuberRegressor"><code class="xref py py-class docutils literal notranslate"><span class="pre">linear_model.HuberRegressor</span></code></a> that was
broken when <code class="docutils literal notranslate"><span class="pre">X</span></code> was of dtype bool. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13328">#13328</a> by <a class="reference external" href="http://alexandre.gramfort.net">Alexandre Gramfort</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> Fixed a performance issue of <code class="docutils literal notranslate"><span class="pre">saga</span></code> and <code class="docutils literal notranslate"><span class="pre">sag</span></code> solvers when called
in a <a class="reference external" href="https://joblib.readthedocs.io/en/latest/generated/joblib.Parallel.html#joblib.Parallel" title="(in joblib v1.3.0.dev0)"><code class="xref py py-class docutils literal notranslate"><span class="pre">joblib.Parallel</span></code></a> setting with <code class="docutils literal notranslate"><span class="pre">n_jobs</span> <span class="pre">></span> <span class="pre">1</span></code> and
<code class="docutils literal notranslate"><span class="pre">backend="threading"</span></code>, causing them to perform worse than in the sequential
case. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13389">#13389</a> by <a class="reference external" href="https://github.com/pierreglaser">Pierre Glaser</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> Fixed a bug in
<code class="xref py py-class docutils literal notranslate"><span class="pre">linear_model.stochastic_gradient.BaseSGDClassifier</span></code> that was not
deterministic when trained in a multi-class setting on several threads.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13422">#13422</a> by <a class="reference external" href="https://github.com/ClemDoum">Clément Doumouro</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> Fixed bug in <a class="reference internal" href="../modules/generated/sklearn.linear_model.ridge_regression.html#sklearn.linear_model.ridge_regression" title="sklearn.linear_model.ridge_regression"><code class="xref py py-func docutils literal notranslate"><span class="pre">linear_model.ridge_regression</span></code></a>,
<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> and
<a class="reference internal" href="../modules/generated/sklearn.linear_model.RidgeClassifier.html#sklearn.linear_model.RidgeClassifier" title="sklearn.linear_model.RidgeClassifier"><code class="xref py py-class docutils literal notranslate"><span class="pre">linear_model.RidgeClassifier</span></code></a> that
caused unhandled exception for arguments <code class="docutils literal notranslate"><span class="pre">return_intercept=True</span></code> and
<code class="docutils literal notranslate"><span class="pre">solver=auto</span></code> (default) or any other solver different from <code class="docutils literal notranslate"><span class="pre">sag</span></code>.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13363">#13363</a> by <a class="reference external" href="https://github.com/btel">Bartosz Telenczuk</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_regression.html#sklearn.linear_model.ridge_regression" title="sklearn.linear_model.ridge_regression"><code class="xref py py-func docutils literal notranslate"><span class="pre">linear_model.ridge_regression</span></code></a> will now raise an exception
if <code class="docutils literal notranslate"><span class="pre">return_intercept=True</span></code> and solver is different from <code class="docutils literal notranslate"><span class="pre">sag</span></code>. Previously,
only warning was issued. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13363">#13363</a> by <a class="reference external" href="https://github.com/btel">Bartosz Telenczuk</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_regression.html#sklearn.linear_model.ridge_regression" title="sklearn.linear_model.ridge_regression"><code class="xref py py-func docutils literal notranslate"><span class="pre">linear_model.ridge_regression</span></code></a> will choose <code class="docutils literal notranslate"><span class="pre">sparse_cg</span></code>
solver for sparse inputs when <code class="docutils literal notranslate"><span class="pre">solver=auto</span></code> and <code class="docutils literal notranslate"><span class="pre">sample_weight</span></code>
is provided (previously <code class="docutils literal notranslate"><span class="pre">cholesky</span></code> solver was selected).
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13363">#13363</a> by <a class="reference external" href="https://github.com/btel">Bartosz Telenczuk</a></p></li>
<li><p><span class="raw-html"><span class="badge badge-warning">API Change</span></span> The use of <a class="reference internal" href="../modules/generated/sklearn.linear_model.lars_path.html#sklearn.linear_model.lars_path" title="sklearn.linear_model.lars_path"><code class="xref py py-class docutils literal notranslate"><span class="pre">linear_model.lars_path</span></code></a> with <code class="docutils literal notranslate"><span class="pre">X=None</span></code>
while passing <code class="docutils literal notranslate"><span class="pre">Gram</span></code> is deprecated in version 0.21 and will be removed
in version 0.23. Use <a class="reference internal" href="../modules/generated/sklearn.linear_model.lars_path_gram.html#sklearn.linear_model.lars_path_gram" title="sklearn.linear_model.lars_path_gram"><code class="xref py py-class docutils literal notranslate"><span class="pre">linear_model.lars_path_gram</span></code></a> instead.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/11699">#11699</a> by <a class="reference external" href="https://github.com/yukuairoy">Kuai Yu</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-warning">API Change</span></span> <code class="xref py py-func docutils literal notranslate"><span class="pre">linear_model.logistic_regression_path</span></code> is deprecated
in version 0.21 and will be removed in version 0.23.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/12821">#12821</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.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> with leave-one-out cross-validation
now correctly fits an intercept when <code class="docutils literal notranslate"><span class="pre">fit_intercept=True</span></code> and the design
matrix is sparse. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/issues/13350">#13350</a> by <a class="reference external" href="https://github.com/jeromedockes">Jérôme Dockès</a></p></li>
</ul>
</section>
<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 heading">¶</a></h3>
<ul class="simple">
<li><p><span class="raw-html"><span class="badge badge-info">Efficiency</span></span> Make <code class="xref py py-func docutils literal notranslate"><span class="pre">manifold.tsne.trustworthiness</span></code> use an inverted index
instead of an <code class="docutils literal notranslate"><span class="pre">np.where</span></code> lookup to find the rank of neighbors in the input
space. This improves efficiency in particular when computed with
lots of neighbors and/or small datasets.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/9907">#9907</a> by <a class="reference external" href="https://github.com/wdevazelhes">William de Vazelhes</a>.</p></li>
</ul>
</section>
<section id="id16">
<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="#id16" title="Permalink to this heading">¶</a></h3>
<ul class="simple">
<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.max_error.html#sklearn.metrics.max_error" title="sklearn.metrics.max_error"><code class="xref py py-func docutils literal notranslate"><span class="pre">metrics.max_error</span></code></a> metric and a corresponding
<code class="docutils literal notranslate"><span class="pre">'max_error'</span></code> scorer for single output regression.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/12232">#12232</a> by <a class="reference external" href="https://github.com/whiletruelearn">Krishna Sangeeth</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.multilabel_confusion_matrix.html#sklearn.metrics.multilabel_confusion_matrix" title="sklearn.metrics.multilabel_confusion_matrix"><code class="xref py py-func docutils literal notranslate"><span class="pre">metrics.multilabel_confusion_matrix</span></code></a>, which calculates a
confusion matrix with true positive, false positive, false negative and true
negative counts for each class. This facilitates the calculation of set-wise
metrics such as recall, specificity, fall out and miss rate.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/11179">#11179</a> by <a class="reference external" href="https://github.com/ShangwuYao">Shangwu Yao</a> and <a class="reference external" href="https://joelnothman.com/">Joel Nothman</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.jaccard_score.html#sklearn.metrics.jaccard_score" title="sklearn.metrics.jaccard_score"><code class="xref py py-func docutils literal notranslate"><span class="pre">metrics.jaccard_score</span></code></a> has been added to calculate the
Jaccard coefficient as an evaluation metric for binary, multilabel and
multiclass tasks, with an interface analogous to <a class="reference internal" href="../modules/generated/sklearn.metrics.f1_score.html#sklearn.metrics.f1_score" title="sklearn.metrics.f1_score"><code class="xref py py-func docutils literal notranslate"><span class="pre">metrics.f1_score</span></code></a>.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13151">#13151</a> by <a class="reference external" href="https://github.com/gxyd">Gaurav Dhingra</a> and <a class="reference external" href="https://joelnothman.com/">Joel Nothman</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-success">Feature</span></span> Added <a class="reference internal" href="../modules/generated/sklearn.metrics.pairwise.haversine_distances.html#sklearn.metrics.pairwise.haversine_distances" title="sklearn.metrics.pairwise.haversine_distances"><code class="xref py py-func docutils literal notranslate"><span class="pre">metrics.pairwise.haversine_distances</span></code></a> which can be
accessed with <code class="docutils literal notranslate"><span class="pre">metric='pairwise'</span></code> through <a class="reference internal" href="../modules/generated/sklearn.metrics.pairwise_distances.html#sklearn.metrics.pairwise_distances" title="sklearn.metrics.pairwise_distances"><code class="xref py py-func docutils literal notranslate"><span class="pre">metrics.pairwise_distances</span></code></a>
and estimators. (Haversine distance was previously available for nearest
neighbors calculation.) <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/12568">#12568</a> by <a class="reference external" href="https://github.com/xuewei4d">Wei Xue</a>,
<a class="reference external" href="https://github.com/eamanu">Emmanuel Arias</a> and <a class="reference external" href="https://joelnothman.com/">Joel Nothman</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-info">Efficiency</span></span> Faster <a class="reference internal" href="../modules/generated/sklearn.metrics.pairwise_distances.html#sklearn.metrics.pairwise_distances" title="sklearn.metrics.pairwise_distances"><code class="xref py py-func docutils literal notranslate"><span class="pre">metrics.pairwise_distances</span></code></a> with <code class="docutils literal notranslate"><span class="pre">n_jobs</span></code>
> 1 by using a thread-based backend, instead of process-based backends.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/8216">#8216</a> by <a class="reference external" href="https://github.com/pierreglaser">Pierre Glaser</a> and
<a class="reference external" href="https://github.com/zanospi">Romuald Menuet</a></p></li>
<li><p><span class="raw-html"><span class="badge badge-info">Efficiency</span></span> The pairwise manhattan distances with sparse input now uses the
BLAS shipped with scipy instead of the bundled BLAS. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/12732">#12732</a> by
<a class="reference external" href="https://github.com/jeremiedbb">Jérémie du Boisberranger</a></p></li>
<li><p><span class="raw-html"><span class="badge badge-info">Enhancement</span></span> Use label <code class="docutils literal notranslate"><span class="pre">accuracy</span></code> instead of <code class="docutils literal notranslate"><span class="pre">micro-average</span></code> on
<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> to avoid confusion. <code class="docutils literal notranslate"><span class="pre">micro-average</span></code> is
only shown for multi-label or multi-class with a subset of classes because
it is otherwise identical to accuracy.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/12334">#12334</a> by <a class="reference external" href="https://github.com/eamanu@eamanu.com">Emmanuel Arias</a>,
<a class="reference external" href="https://joelnothman.com/">Joel Nothman</a> and <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> Added <code class="docutils literal notranslate"><span class="pre">beta</span></code> parameter to
<a class="reference internal" href="../modules/generated/sklearn.metrics.homogeneity_completeness_v_measure.html#sklearn.metrics.homogeneity_completeness_v_measure" title="sklearn.metrics.homogeneity_completeness_v_measure"><code class="xref py py-func docutils literal notranslate"><span class="pre">metrics.homogeneity_completeness_v_measure</span></code></a> and
<a class="reference internal" href="../modules/generated/sklearn.metrics.v_measure_score.html#sklearn.metrics.v_measure_score" title="sklearn.metrics.v_measure_score"><code class="xref py py-func docutils literal notranslate"><span class="pre">metrics.v_measure_score</span></code></a> to configure the
tradeoff between homogeneity and completeness.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13607">#13607</a> by <a class="reference external" href="https://github.com/scouvreur">Stephane Couvreur</a> and
and <a class="reference external" href="https://github.com/ivsanro1">Ivan Sanchez</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> The metric <a class="reference internal" href="../modules/generated/sklearn.metrics.r2_score.html#sklearn.metrics.r2_score" title="sklearn.metrics.r2_score"><code class="xref py py-func docutils literal notranslate"><span class="pre">metrics.r2_score</span></code></a> is degenerate with a single sample
and now it returns NaN and raises <a class="reference internal" href="../modules/generated/sklearn.exceptions.UndefinedMetricWarning.html#sklearn.exceptions.UndefinedMetricWarning" title="sklearn.exceptions.UndefinedMetricWarning"><code class="xref py py-class docutils literal notranslate"><span class="pre">exceptions.UndefinedMetricWarning</span></code></a>.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/12855">#12855</a> by <a class="reference external" href="https://github.com/psendyk">Pawel Sendyk</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.metrics.brier_score_loss.html#sklearn.metrics.brier_score_loss" title="sklearn.metrics.brier_score_loss"><code class="xref py py-func docutils literal notranslate"><span class="pre">metrics.brier_score_loss</span></code></a> will sometimes
return incorrect result when there’s only one class in <code class="docutils literal notranslate"><span class="pre">y_true</span></code>.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13628">#13628</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> Fixed a bug in <a class="reference internal" href="../modules/generated/sklearn.metrics.label_ranking_average_precision_score.html#sklearn.metrics.label_ranking_average_precision_score" title="sklearn.metrics.label_ranking_average_precision_score"><code class="xref py py-func docutils literal notranslate"><span class="pre">metrics.label_ranking_average_precision_score</span></code></a>
where sample_weight wasn’t taken into account for samples with degenerate
labels.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13447">#13447</a> by <a class="reference external" href="https://github.com/dpwe">Dan Ellis</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-warning">API Change</span></span> The parameter <code class="docutils literal notranslate"><span class="pre">labels</span></code> in <a class="reference internal" href="../modules/generated/sklearn.metrics.hamming_loss.html#sklearn.metrics.hamming_loss" title="sklearn.metrics.hamming_loss"><code class="xref py py-func docutils literal notranslate"><span class="pre">metrics.hamming_loss</span></code></a> is deprecated
in version 0.21 and will be removed in version 0.23. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/10580">#10580</a> by
<a class="reference external" href="https://github.com/reshamas">Reshama Shaikh</a> and <a class="reference external" href="https://github.com/SandraMNE">Sandra Mitrovic</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> The function <a class="reference internal" href="../modules/generated/sklearn.metrics.pairwise.euclidean_distances.html#sklearn.metrics.pairwise.euclidean_distances" title="sklearn.metrics.pairwise.euclidean_distances"><code class="xref py py-func docutils literal notranslate"><span class="pre">metrics.pairwise.euclidean_distances</span></code></a>, and
therefore several estimators with <code class="docutils literal notranslate"><span class="pre">metric='euclidean'</span></code>, suffered from
numerical precision issues with <code class="docutils literal notranslate"><span class="pre">float32</span></code> features. Precision has been
increased at the cost of a small drop of performance. <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13554">#13554</a> by
<a class="reference external" href="https://github.com/Celelibi">@Celelibi</a> and <a class="reference external" href="https://github.com/jeremiedbb">Jérémie du Boisberranger</a>.</p></li>
<li><p><span class="raw-html"><span class="badge badge-warning">API Change</span></span> <code class="xref py py-func docutils literal notranslate"><span class="pre">metrics.jaccard_similarity_score</span></code> is deprecated in favour of
the more consistent <a class="reference internal" href="../modules/generated/sklearn.metrics.jaccard_score.html#sklearn.metrics.jaccard_score" title="sklearn.metrics.jaccard_score"><code class="xref py py-func docutils literal notranslate"><span class="pre">metrics.jaccard_score</span></code></a>. The former behavior for
binary and multiclass targets is broken.
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/pull/13151">#13151</a> by <a class="reference external" href="https://joelnothman.com/">Joel Nothman</a>.</p></li>
</ul>
</section>
<section id="sklearn-mixture">
<h3><a class="reference internal" href="../modules/classes.html#module-sklearn.mixture" title="sklearn.mixture"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.mixture</span></code></a><a class="headerlink" href="#sklearn-mixture" title="Permalink to this heading">¶</a></h3>
<ul class="simple">
<li><p><span class="raw-html"><span class="badge badge-danger">Fix</span></span> Fixed a bug in <code class="xref py py-class docutils literal notranslate"><span class="pre">mixture.BaseMixture</span></code> and therefore on estimators
based on it, i.e. <a class="reference internal" href="../modules/generated/sklearn.mixture.GaussianMixture.html#sklearn.mixture.GaussianMixture" title="sklearn.mixture.GaussianMixture"><code class="xref py py-class docutils literal notranslate"><span class="pre">mixture.GaussianMixture</span></code></a> and
<a class="reference internal" href="../modules/generated/sklearn.mixture.BayesianGaussianMixture.html#sklearn.mixture.BayesianGaussianMixture" title="sklearn.mixture.BayesianGaussianMixture"><code class="xref py py-class docutils literal notranslate"><span class="pre">mixture.BayesianGaussianMixture</span></code></a>, where <code class="docutils literal notranslate"><span class="pre">fit_predict</span></code> and