-
Notifications
You must be signed in to change notification settings - Fork 81
/
Copy pathfeature_selection.html
1113 lines (883 loc) · 73.7 KB
/
feature_selection.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>
<html lang="en" data-content_root="../" >
<head>
<meta charset="utf-8" />
<meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="viewport" content="width=device-width, initial-scale=1" />
<meta property="og:title" content="1.13. Feature selection" />
<meta property="og:type" content="website" />
<meta property="og:url" content="https://scikit-learn/stable/modules/feature_selection.html" />
<meta property="og:site_name" content="scikit-learn" />
<meta property="og:description" content="The classes in the sklearn.feature_selection module can be used for feature selection/dimensionality reduction on sample sets, either to improve estimators’ accuracy scores or to boost their perfor..." />
<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="The classes in the sklearn.feature_selection module can be used for feature selection/dimensionality reduction on sample sets, either to improve estimators’ accuracy scores or to boost their perfor..." />
<title>1.13. Feature selection — scikit-learn 1.7.dev0 documentation</title>
<script data-cfasync="false">
document.documentElement.dataset.mode = localStorage.getItem("mode") || "";
document.documentElement.dataset.theme = localStorage.getItem("theme") || "";
</script>
<!--
this give us a css class that will be invisible only if js is disabled
-->
<noscript>
<style>
.pst-js-only { display: none !important; }
</style>
</noscript>
<!-- Loaded before other Sphinx assets -->
<link href="../_static/styles/theme.css?digest=8878045cc6db502f8baf" rel="stylesheet" />
<link href="../_static/styles/pydata-sphinx-theme.css?digest=8878045cc6db502f8baf" rel="stylesheet" />
<link rel="stylesheet" type="text/css" href="../_static/pygments.css?v=8f2a1f02" />
<link rel="stylesheet" type="text/css" href="../_static/copybutton.css?v=76b2166b" />
<link rel="stylesheet" type="text/css" href="../_static/plot_directive.css" />
<link rel="stylesheet" type="text/css" href="https://fonts.googleapis.com/css?family=Vibur" />
<link rel="stylesheet" type="text/css" href="../_static/jupyterlite_sphinx.css?v=2c9f8f05" />
<link rel="stylesheet" type="text/css" href="../_static/sg_gallery.css?v=d2d258e8" />
<link rel="stylesheet" type="text/css" href="../_static/sg_gallery-binder.css?v=f4aeca0c" />
<link rel="stylesheet" type="text/css" href="../_static/sg_gallery-dataframe.css?v=2082cf3c" />
<link rel="stylesheet" type="text/css" href="../_static/sg_gallery-rendered-html.css?v=1277b6f3" />
<link rel="stylesheet" type="text/css" href="../_static/sphinx-design.min.css?v=95c83b7e" />
<link rel="stylesheet" type="text/css" href="../_static/styles/colors.css?v=cc94ab7d" />
<link rel="stylesheet" type="text/css" href="../_static/styles/custom.css?v=8f525996" />
<!-- So that users can add custom icons -->
<script src="../_static/scripts/fontawesome.js?digest=8878045cc6db502f8baf"></script>
<!-- Pre-loaded scripts that we'll load fully later -->
<link rel="preload" as="script" href="../_static/scripts/bootstrap.js?digest=8878045cc6db502f8baf" />
<link rel="preload" as="script" href="../_static/scripts/pydata-sphinx-theme.js?digest=8878045cc6db502f8baf" />
<script src="../_static/documentation_options.js?v=473747f4"></script>
<script src="../_static/doctools.js?v=9bcbadda"></script>
<script src="../_static/sphinx_highlight.js?v=dc90522c"></script>
<script src="../_static/clipboard.min.js?v=a7894cd8"></script>
<script src="../_static/copybutton.js?v=97f0b27d"></script>
<script src="../_static/jupyterlite_sphinx.js?v=96e329c5"></script>
<script src="../_static/design-tabs.js?v=f930bc37"></script>
<script data-domain="scikit-learn.org" defer="defer" src="https://views.scientific-python.org/js/script.js"></script>
<script async="async" src="https://cdn.jsdelivr.net/npm/mathjax@3/es5/tex-chtml.js"></script>
<script>DOCUMENTATION_OPTIONS.pagename = 'modules/feature_selection';</script>
<script>
DOCUMENTATION_OPTIONS.theme_version = '0.16.1';
DOCUMENTATION_OPTIONS.theme_switcher_json_url = 'https://scikit-learn.org/dev/_static/versions.json';
DOCUMENTATION_OPTIONS.theme_switcher_version_match = '1.7.dev0';
DOCUMENTATION_OPTIONS.show_version_warning_banner =
true;
</script>
<script src="../_static/scripts/dropdown.js?v=d6825577"></script>
<script src="../_static/scripts/version-switcher.js?v=a6dd8357"></script>
<script src="../_static/scripts/sg_plotly_resize.js?v=2167d4db"></script>
<link rel="canonical" href="https://scikit-learn.org/stable/modules/feature_selection.html" />
<link rel="icon" href="../_static/favicon.ico"/>
<link rel="author" title="About these documents" href="../about.html" />
<link rel="search" title="Search" href="../search.html" />
<link rel="next" title="1.14. Semi-supervised learning" href="semi_supervised.html" />
<link rel="prev" title="1.12. Multiclass and multioutput algorithms" href="multiclass.html" />
<meta name="viewport" content="width=device-width, initial-scale=1"/>
<meta name="docsearch:language" content="en"/>
<meta name="docsearch:version" content="1.7" />
</head>
<body data-bs-spy="scroll" data-bs-target=".bd-toc-nav" data-offset="180" data-bs-root-margin="0px 0px -60%" data-default-mode="">
<div id="pst-skip-link" class="skip-link d-print-none"><a href="#main-content">Skip to main content</a></div>
<div id="pst-scroll-pixel-helper"></div>
<button type="button" class="btn rounded-pill" id="pst-back-to-top">
<i class="fa-solid fa-arrow-up"></i>Back to top</button>
<dialog id="pst-search-dialog">
<form class="bd-search d-flex align-items-center"
action="../search.html"
method="get">
<i class="fa-solid fa-magnifying-glass"></i>
<input type="search"
class="form-control"
name="q"
placeholder="Search the docs ..."
aria-label="Search the docs ..."
autocomplete="off"
autocorrect="off"
autocapitalize="off"
spellcheck="false"/>
<span class="search-button__kbd-shortcut"><kbd class="kbd-shortcut__modifier">Ctrl</kbd>+<kbd>K</kbd></span>
</form>
</dialog>
<div class="pst-async-banner-revealer d-none">
<aside id="bd-header-version-warning" class="d-none d-print-none" aria-label="Version warning"></aside>
</div>
<header class="bd-header navbar navbar-expand-lg bd-navbar d-print-none">
<div class="bd-header__inner bd-page-width">
<button class="pst-navbar-icon sidebar-toggle primary-toggle" aria-label="Site navigation">
<span class="fa-solid fa-bars"></span>
</button>
<div class=" navbar-header-items__start">
<div class="navbar-item">
<a class="navbar-brand logo" href="../index.html">
<img src="../_static/scikit-learn-logo-small.png" class="logo__image only-light" alt="scikit-learn homepage"/>
<img src="../_static/scikit-learn-logo-small.png" class="logo__image only-dark pst-js-only" alt="scikit-learn homepage"/>
</a></div>
</div>
<div class=" navbar-header-items">
<div class="me-auto navbar-header-items__center">
<div class="navbar-item">
<nav>
<ul class="bd-navbar-elements navbar-nav">
<li class="nav-item ">
<a class="nav-link nav-internal" href="../install.html">
Install
</a>
</li>
<li class="nav-item current active">
<a class="nav-link nav-internal" href="../user_guide.html">
User Guide
</a>
</li>
<li class="nav-item ">
<a class="nav-link nav-internal" href="../api/index.html">
API
</a>
</li>
<li class="nav-item ">
<a class="nav-link nav-internal" href="../auto_examples/index.html">
Examples
</a>
</li>
<li class="nav-item ">
<a class="nav-link nav-external" href="https://blog.scikit-learn.org/">
Community
</a>
</li>
<li class="nav-item dropdown">
<button class="btn dropdown-toggle nav-item" type="button"
data-bs-toggle="dropdown" aria-expanded="false"
aria-controls="pst-nav-more-links">
More
</button>
<ul id="pst-nav-more-links" class="dropdown-menu">
<li class=" ">
<a class="nav-link dropdown-item nav-internal" href="../getting_started.html">
Getting Started
</a>
</li>
<li class=" ">
<a class="nav-link dropdown-item nav-internal" href="../whats_new.html">
Release History
</a>
</li>
<li class=" ">
<a class="nav-link dropdown-item nav-internal" href="../glossary.html">
Glossary
</a>
</li>
<li class=" ">
<a class="nav-link dropdown-item nav-internal" href="../developers/index.html">
Development
</a>
</li>
<li class=" ">
<a class="nav-link dropdown-item nav-internal" href="../faq.html">
FAQ
</a>
</li>
<li class=" ">
<a class="nav-link dropdown-item nav-internal" href="../support.html">
Support
</a>
</li>
<li class=" ">
<a class="nav-link dropdown-item nav-internal" href="../related_projects.html">
Related Projects
</a>
</li>
<li class=" ">
<a class="nav-link dropdown-item nav-internal" href="../roadmap.html">
Roadmap
</a>
</li>
<li class=" ">
<a class="nav-link dropdown-item nav-internal" href="../governance.html">
Governance
</a>
</li>
<li class=" ">
<a class="nav-link dropdown-item nav-internal" href="../about.html">
About us
</a>
</li>
</ul>
</li>
</ul>
</nav></div>
</div>
<div class="navbar-header-items__end">
<div class="navbar-item navbar-persistent--container">
<button class="btn btn-sm pst-navbar-icon search-button search-button__button pst-js-only" title="Search" aria-label="Search" data-bs-placement="bottom" data-bs-toggle="tooltip">
<i class="fa-solid fa-magnifying-glass fa-lg"></i>
</button>
</div>
<div class="navbar-item">
<button class="btn btn-sm nav-link pst-navbar-icon theme-switch-button pst-js-only" aria-label="Color mode" data-bs-title="Color mode" data-bs-placement="bottom" data-bs-toggle="tooltip">
<i class="theme-switch fa-solid fa-sun fa-lg" data-mode="light" title="Light"></i>
<i class="theme-switch fa-solid fa-moon fa-lg" data-mode="dark" title="Dark"></i>
<i class="theme-switch fa-solid fa-circle-half-stroke fa-lg" data-mode="auto" title="System Settings"></i>
</button></div>
<div class="navbar-item"><ul class="navbar-icon-links"
aria-label="Icon Links">
<li class="nav-item">
<a href="https://github.com/scikit-learn/scikit-learn" title="GitHub" class="nav-link pst-navbar-icon" rel="noopener" target="_blank" data-bs-toggle="tooltip" data-bs-placement="bottom"><i class="fa-brands fa-square-github fa-lg" aria-hidden="true"></i>
<span class="sr-only">GitHub</span></a>
</li>
</ul></div>
<div class="navbar-item">
<div class="version-switcher__container dropdown pst-js-only">
<button id="pst-version-switcher-button-2"
type="button"
class="version-switcher__button btn btn-sm dropdown-toggle"
data-bs-toggle="dropdown"
aria-haspopup="listbox"
aria-controls="pst-version-switcher-list-2"
aria-label="Version switcher list"
>
Choose version <!-- this text may get changed later by javascript -->
<span class="caret"></span>
</button>
<div id="pst-version-switcher-list-2"
class="version-switcher__menu dropdown-menu list-group-flush py-0"
role="listbox" aria-labelledby="pst-version-switcher-button-2">
<!-- dropdown will be populated by javascript on page load -->
</div>
</div></div>
</div>
</div>
<div class="navbar-persistent--mobile">
<button class="btn btn-sm pst-navbar-icon search-button search-button__button pst-js-only" title="Search" aria-label="Search" data-bs-placement="bottom" data-bs-toggle="tooltip">
<i class="fa-solid fa-magnifying-glass fa-lg"></i>
</button>
</div>
<button class="pst-navbar-icon sidebar-toggle secondary-toggle" aria-label="On this page">
<span class="fa-solid fa-outdent"></span>
</button>
</div>
</header>
<div class="bd-container">
<div class="bd-container__inner bd-page-width">
<dialog id="pst-primary-sidebar-modal"></dialog>
<div id="pst-primary-sidebar" class="bd-sidebar-primary bd-sidebar">
<div class="sidebar-header-items sidebar-primary__section">
<div class="sidebar-header-items__center">
<div class="navbar-item">
<nav>
<ul class="bd-navbar-elements navbar-nav">
<li class="nav-item ">
<a class="nav-link nav-internal" href="../install.html">
Install
</a>
</li>
<li class="nav-item current active">
<a class="nav-link nav-internal" href="../user_guide.html">
User Guide
</a>
</li>
<li class="nav-item ">
<a class="nav-link nav-internal" href="../api/index.html">
API
</a>
</li>
<li class="nav-item ">
<a class="nav-link nav-internal" href="../auto_examples/index.html">
Examples
</a>
</li>
<li class="nav-item ">
<a class="nav-link nav-external" href="https://blog.scikit-learn.org/">
Community
</a>
</li>
<li class="nav-item ">
<a class="nav-link nav-internal" href="../getting_started.html">
Getting Started
</a>
</li>
<li class="nav-item ">
<a class="nav-link nav-internal" href="../whats_new.html">
Release History
</a>
</li>
<li class="nav-item ">
<a class="nav-link nav-internal" href="../glossary.html">
Glossary
</a>
</li>
<li class="nav-item ">
<a class="nav-link nav-internal" href="../developers/index.html">
Development
</a>
</li>
<li class="nav-item ">
<a class="nav-link nav-internal" href="../faq.html">
FAQ
</a>
</li>
<li class="nav-item ">
<a class="nav-link nav-internal" href="../support.html">
Support
</a>
</li>
<li class="nav-item ">
<a class="nav-link nav-internal" href="../related_projects.html">
Related Projects
</a>
</li>
<li class="nav-item ">
<a class="nav-link nav-internal" href="../roadmap.html">
Roadmap
</a>
</li>
<li class="nav-item ">
<a class="nav-link nav-internal" href="../governance.html">
Governance
</a>
</li>
<li class="nav-item ">
<a class="nav-link nav-internal" href="../about.html">
About us
</a>
</li>
</ul>
</nav></div>
</div>
<div class="sidebar-header-items__end">
<div class="navbar-item">
<button class="btn btn-sm nav-link pst-navbar-icon theme-switch-button pst-js-only" aria-label="Color mode" data-bs-title="Color mode" data-bs-placement="bottom" data-bs-toggle="tooltip">
<i class="theme-switch fa-solid fa-sun fa-lg" data-mode="light" title="Light"></i>
<i class="theme-switch fa-solid fa-moon fa-lg" data-mode="dark" title="Dark"></i>
<i class="theme-switch fa-solid fa-circle-half-stroke fa-lg" data-mode="auto" title="System Settings"></i>
</button></div>
<div class="navbar-item"><ul class="navbar-icon-links"
aria-label="Icon Links">
<li class="nav-item">
<a href="https://github.com/scikit-learn/scikit-learn" title="GitHub" class="nav-link pst-navbar-icon" rel="noopener" target="_blank" data-bs-toggle="tooltip" data-bs-placement="bottom"><i class="fa-brands fa-square-github fa-lg" aria-hidden="true"></i>
<span class="sr-only">GitHub</span></a>
</li>
</ul></div>
<div class="navbar-item">
<div class="version-switcher__container dropdown pst-js-only">
<button id="pst-version-switcher-button-3"
type="button"
class="version-switcher__button btn btn-sm dropdown-toggle"
data-bs-toggle="dropdown"
aria-haspopup="listbox"
aria-controls="pst-version-switcher-list-3"
aria-label="Version switcher list"
>
Choose version <!-- this text may get changed later by javascript -->
<span class="caret"></span>
</button>
<div id="pst-version-switcher-list-3"
class="version-switcher__menu dropdown-menu list-group-flush py-0"
role="listbox" aria-labelledby="pst-version-switcher-button-3">
<!-- dropdown will be populated by javascript on page load -->
</div>
</div></div>
</div>
</div>
<div class="sidebar-primary-items__start sidebar-primary__section">
<div class="sidebar-primary-item">
<nav class="bd-docs-nav bd-links"
aria-label="Section Navigation">
<p class="bd-links__title" role="heading" aria-level="1">Section Navigation</p>
<div class="bd-toc-item navbar-nav"><ul class="current nav bd-sidenav">
<li class="toctree-l1 current active has-children"><a class="reference internal" href="../supervised_learning.html">1. Supervised learning</a><details open="open"><summary><span class="toctree-toggle" role="presentation"><i class="fa-solid fa-chevron-down"></i></span></summary><ul class="current">
<li class="toctree-l2"><a class="reference internal" href="linear_model.html">1.1. Linear Models</a></li>
<li class="toctree-l2"><a class="reference internal" href="lda_qda.html">1.2. Linear and Quadratic Discriminant Analysis</a></li>
<li class="toctree-l2"><a class="reference internal" href="kernel_ridge.html">1.3. Kernel ridge regression</a></li>
<li class="toctree-l2"><a class="reference internal" href="svm.html">1.4. Support Vector Machines</a></li>
<li class="toctree-l2"><a class="reference internal" href="sgd.html">1.5. Stochastic Gradient Descent</a></li>
<li class="toctree-l2"><a class="reference internal" href="neighbors.html">1.6. Nearest Neighbors</a></li>
<li class="toctree-l2"><a class="reference internal" href="gaussian_process.html">1.7. Gaussian Processes</a></li>
<li class="toctree-l2"><a class="reference internal" href="cross_decomposition.html">1.8. Cross decomposition</a></li>
<li class="toctree-l2"><a class="reference internal" href="naive_bayes.html">1.9. Naive Bayes</a></li>
<li class="toctree-l2"><a class="reference internal" href="tree.html">1.10. Decision Trees</a></li>
<li class="toctree-l2"><a class="reference internal" href="ensemble.html">1.11. Ensembles: Gradient boosting, random forests, bagging, voting, stacking</a></li>
<li class="toctree-l2"><a class="reference internal" href="multiclass.html">1.12. Multiclass and multioutput algorithms</a></li>
<li class="toctree-l2 current active"><a class="current reference internal" href="#">1.13. Feature selection</a></li>
<li class="toctree-l2"><a class="reference internal" href="semi_supervised.html">1.14. Semi-supervised learning</a></li>
<li class="toctree-l2"><a class="reference internal" href="isotonic.html">1.15. Isotonic regression</a></li>
<li class="toctree-l2"><a class="reference internal" href="calibration.html">1.16. Probability calibration</a></li>
<li class="toctree-l2"><a class="reference internal" href="neural_networks_supervised.html">1.17. Neural network models (supervised)</a></li>
</ul>
</details></li>
<li class="toctree-l1 has-children"><a class="reference internal" href="../unsupervised_learning.html">2. Unsupervised learning</a><details><summary><span class="toctree-toggle" role="presentation"><i class="fa-solid fa-chevron-down"></i></span></summary><ul>
<li class="toctree-l2"><a class="reference internal" href="mixture.html">2.1. Gaussian mixture models</a></li>
<li class="toctree-l2"><a class="reference internal" href="manifold.html">2.2. Manifold learning</a></li>
<li class="toctree-l2"><a class="reference internal" href="clustering.html">2.3. Clustering</a></li>
<li class="toctree-l2"><a class="reference internal" href="biclustering.html">2.4. Biclustering</a></li>
<li class="toctree-l2"><a class="reference internal" href="decomposition.html">2.5. Decomposing signals in components (matrix factorization problems)</a></li>
<li class="toctree-l2"><a class="reference internal" href="covariance.html">2.6. Covariance estimation</a></li>
<li class="toctree-l2"><a class="reference internal" href="outlier_detection.html">2.7. Novelty and Outlier Detection</a></li>
<li class="toctree-l2"><a class="reference internal" href="density.html">2.8. Density Estimation</a></li>
<li class="toctree-l2"><a class="reference internal" href="neural_networks_unsupervised.html">2.9. Neural network models (unsupervised)</a></li>
</ul>
</details></li>
<li class="toctree-l1 has-children"><a class="reference internal" href="../model_selection.html">3. Model selection and evaluation</a><details><summary><span class="toctree-toggle" role="presentation"><i class="fa-solid fa-chevron-down"></i></span></summary><ul>
<li class="toctree-l2"><a class="reference internal" href="cross_validation.html">3.1. Cross-validation: evaluating estimator performance</a></li>
<li class="toctree-l2"><a class="reference internal" href="grid_search.html">3.2. Tuning the hyper-parameters of an estimator</a></li>
<li class="toctree-l2"><a class="reference internal" href="classification_threshold.html">3.3. Tuning the decision threshold for class prediction</a></li>
<li class="toctree-l2"><a class="reference internal" href="model_evaluation.html">3.4. Metrics and scoring: quantifying the quality of predictions</a></li>
<li class="toctree-l2"><a class="reference internal" href="learning_curve.html">3.5. Validation curves: plotting scores to evaluate models</a></li>
</ul>
</details></li>
<li class="toctree-l1"><a class="reference internal" href="../metadata_routing.html">4. Metadata Routing</a></li>
<li class="toctree-l1 has-children"><a class="reference internal" href="../inspection.html">5. Inspection</a><details><summary><span class="toctree-toggle" role="presentation"><i class="fa-solid fa-chevron-down"></i></span></summary><ul>
<li class="toctree-l2"><a class="reference internal" href="partial_dependence.html">5.1. Partial Dependence and Individual Conditional Expectation plots</a></li>
<li class="toctree-l2"><a class="reference internal" href="permutation_importance.html">5.2. Permutation feature importance</a></li>
</ul>
</details></li>
<li class="toctree-l1"><a class="reference internal" href="../visualizations.html">6. Visualizations</a></li>
<li class="toctree-l1 has-children"><a class="reference internal" href="../data_transforms.html">7. Dataset transformations</a><details><summary><span class="toctree-toggle" role="presentation"><i class="fa-solid fa-chevron-down"></i></span></summary><ul>
<li class="toctree-l2"><a class="reference internal" href="compose.html">7.1. Pipelines and composite estimators</a></li>
<li class="toctree-l2"><a class="reference internal" href="feature_extraction.html">7.2. Feature extraction</a></li>
<li class="toctree-l2"><a class="reference internal" href="preprocessing.html">7.3. Preprocessing data</a></li>
<li class="toctree-l2"><a class="reference internal" href="impute.html">7.4. Imputation of missing values</a></li>
<li class="toctree-l2"><a class="reference internal" href="unsupervised_reduction.html">7.5. Unsupervised dimensionality reduction</a></li>
<li class="toctree-l2"><a class="reference internal" href="random_projection.html">7.6. Random Projection</a></li>
<li class="toctree-l2"><a class="reference internal" href="kernel_approximation.html">7.7. Kernel Approximation</a></li>
<li class="toctree-l2"><a class="reference internal" href="metrics.html">7.8. Pairwise metrics, Affinities and Kernels</a></li>
<li class="toctree-l2"><a class="reference internal" href="preprocessing_targets.html">7.9. Transforming the prediction target (<code class="docutils literal notranslate"><span class="pre">y</span></code>)</a></li>
</ul>
</details></li>
<li class="toctree-l1 has-children"><a class="reference internal" href="../datasets.html">8. Dataset loading utilities</a><details><summary><span class="toctree-toggle" role="presentation"><i class="fa-solid fa-chevron-down"></i></span></summary><ul>
<li class="toctree-l2"><a class="reference internal" href="../datasets/toy_dataset.html">8.1. Toy datasets</a></li>
<li class="toctree-l2"><a class="reference internal" href="../datasets/real_world.html">8.2. Real world datasets</a></li>
<li class="toctree-l2"><a class="reference internal" href="../datasets/sample_generators.html">8.3. Generated datasets</a></li>
<li class="toctree-l2"><a class="reference internal" href="../datasets/loading_other_datasets.html">8.4. Loading other datasets</a></li>
</ul>
</details></li>
<li class="toctree-l1 has-children"><a class="reference internal" href="../computing.html">9. Computing with scikit-learn</a><details><summary><span class="toctree-toggle" role="presentation"><i class="fa-solid fa-chevron-down"></i></span></summary><ul>
<li class="toctree-l2"><a class="reference internal" href="../computing/scaling_strategies.html">9.1. Strategies to scale computationally: bigger data</a></li>
<li class="toctree-l2"><a class="reference internal" href="../computing/computational_performance.html">9.2. Computational Performance</a></li>
<li class="toctree-l2"><a class="reference internal" href="../computing/parallelism.html">9.3. Parallelism, resource management, and configuration</a></li>
</ul>
</details></li>
<li class="toctree-l1"><a class="reference internal" href="../model_persistence.html">10. Model persistence</a></li>
<li class="toctree-l1"><a class="reference internal" href="../common_pitfalls.html">11. Common pitfalls and recommended practices</a></li>
<li class="toctree-l1 has-children"><a class="reference internal" href="../dispatching.html">12. Dispatching</a><details><summary><span class="toctree-toggle" role="presentation"><i class="fa-solid fa-chevron-down"></i></span></summary><ul>
<li class="toctree-l2"><a class="reference internal" href="array_api.html">12.1. Array API support (experimental)</a></li>
</ul>
</details></li>
<li class="toctree-l1"><a class="reference internal" href="../machine_learning_map.html">13. Choosing the right estimator</a></li>
<li class="toctree-l1"><a class="reference internal" href="../presentations.html">14. External Resources, Videos and Talks</a></li>
</ul>
</div>
</nav></div>
</div>
<div class="sidebar-primary-items__end sidebar-primary__section">
</div>
</div>
<main id="main-content" class="bd-main" role="main">
<div class="bd-content">
<div class="bd-article-container">
<div class="bd-header-article d-print-none">
<div class="header-article-items header-article__inner">
<div class="header-article-items__start">
<div class="header-article-item">
<nav aria-label="Breadcrumb" class="d-print-none">
<ul class="bd-breadcrumbs">
<li class="breadcrumb-item breadcrumb-home">
<a href="../index.html" class="nav-link" aria-label="Home">
<i class="fa-solid fa-home"></i>
</a>
</li>
<li class="breadcrumb-item"><a href="../user_guide.html" class="nav-link">User Guide</a></li>
<li class="breadcrumb-item"><a href="../supervised_learning.html" class="nav-link"><span class="section-number">1. </span>Supervised learning</a></li>
<li class="breadcrumb-item active" aria-current="page"><span class="ellipsis"><span class="section-number">1.13. </span>Feature selection</span></li>
</ul>
</nav>
</div>
</div>
</div>
</div>
<div id="searchbox"></div>
<article class="bd-article">
<section id="feature-selection">
<span id="id1"></span><h1><span class="section-number">1.13. </span>Feature selection<a class="headerlink" href="#feature-selection" title="Link to this heading">#</a></h1>
<p>The classes in the <a class="reference internal" href="../api/sklearn.feature_selection.html#module-sklearn.feature_selection" title="sklearn.feature_selection"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.feature_selection</span></code></a> module can be used
for feature selection/dimensionality reduction on sample sets, either to
improve estimators’ accuracy scores or to boost their performance on very
high-dimensional datasets.</p>
<section id="removing-features-with-low-variance">
<span id="variance-threshold"></span><h2><span class="section-number">1.13.1. </span>Removing features with low variance<a class="headerlink" href="#removing-features-with-low-variance" title="Link to this heading">#</a></h2>
<p><a class="reference internal" href="generated/sklearn.feature_selection.VarianceThreshold.html#sklearn.feature_selection.VarianceThreshold" title="sklearn.feature_selection.VarianceThreshold"><code class="xref py py-class docutils literal notranslate"><span class="pre">VarianceThreshold</span></code></a> is a simple baseline approach to feature selection.
It removes all features whose variance doesn’t meet some threshold.
By default, it removes all zero-variance features,
i.e. features that have the same value in all samples.</p>
<p>As an example, suppose that we have a dataset with boolean features,
and we want to remove all features that are either one or zero (on or off)
in more than 80% of the samples.
Boolean features are Bernoulli random variables,
and the variance of such variables is given by</p>
<div class="math notranslate nohighlight">
\[\mathrm{Var}[X] = p(1 - p)\]</div>
<p>so we can select using the threshold <code class="docutils literal notranslate"><span class="pre">.8</span> <span class="pre">*</span> <span class="pre">(1</span> <span class="pre">-</span> <span class="pre">.8)</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="w"> </span><span class="nn">sklearn.feature_selection</span><span class="w"> </span><span class="kn">import</span> <span class="n">VarianceThreshold</span>
<span class="gp">>>> </span><span class="n">X</span> <span class="o">=</span> <span class="p">[[</span><span class="mi">0</span><span class="p">,</span> <span class="mi">0</span><span class="p">,</span> <span class="mi">1</span><span class="p">],</span> <span class="p">[</span><span class="mi">0</span><span class="p">,</span> <span class="mi">1</span><span class="p">,</span> <span class="mi">0</span><span class="p">],</span> <span class="p">[</span><span class="mi">1</span><span class="p">,</span> <span class="mi">0</span><span class="p">,</span> <span class="mi">0</span><span class="p">],</span> <span class="p">[</span><span class="mi">0</span><span class="p">,</span> <span class="mi">1</span><span class="p">,</span> <span class="mi">1</span><span class="p">],</span> <span class="p">[</span><span class="mi">0</span><span class="p">,</span> <span class="mi">1</span><span class="p">,</span> <span class="mi">0</span><span class="p">],</span> <span class="p">[</span><span class="mi">0</span><span class="p">,</span> <span class="mi">1</span><span class="p">,</span> <span class="mi">1</span><span class="p">]]</span>
<span class="gp">>>> </span><span class="n">sel</span> <span class="o">=</span> <span class="n">VarianceThreshold</span><span class="p">(</span><span class="n">threshold</span><span class="o">=</span><span class="p">(</span><span class="mf">.8</span> <span class="o">*</span> <span class="p">(</span><span class="mi">1</span> <span class="o">-</span> <span class="mf">.8</span><span class="p">)))</span>
<span class="gp">>>> </span><span class="n">sel</span><span class="o">.</span><span class="n">fit_transform</span><span class="p">(</span><span class="n">X</span><span class="p">)</span>
<span class="go">array([[0, 1],</span>
<span class="go"> [1, 0],</span>
<span class="go"> [0, 0],</span>
<span class="go"> [1, 1],</span>
<span class="go"> [1, 0],</span>
<span class="go"> [1, 1]])</span>
</pre></div>
</div>
<p>As expected, <code class="docutils literal notranslate"><span class="pre">VarianceThreshold</span></code> has removed the first column,
which has a probability <span class="math notranslate nohighlight">\(p = 5/6 > .8\)</span> of containing a zero.</p>
</section>
<section id="univariate-feature-selection">
<span id="id2"></span><h2><span class="section-number">1.13.2. </span>Univariate feature selection<a class="headerlink" href="#univariate-feature-selection" title="Link to this heading">#</a></h2>
<p>Univariate feature selection works by selecting the best features based on
univariate statistical tests. It can be seen as a preprocessing step
to an estimator. Scikit-learn exposes feature selection routines
as objects that implement the <code class="docutils literal notranslate"><span class="pre">transform</span></code> method:</p>
<ul class="simple">
<li><p><a class="reference internal" href="generated/sklearn.feature_selection.SelectKBest.html#sklearn.feature_selection.SelectKBest" title="sklearn.feature_selection.SelectKBest"><code class="xref py py-class docutils literal notranslate"><span class="pre">SelectKBest</span></code></a> removes all but the <span class="math notranslate nohighlight">\(k\)</span> highest scoring features</p></li>
<li><p><a class="reference internal" href="generated/sklearn.feature_selection.SelectPercentile.html#sklearn.feature_selection.SelectPercentile" title="sklearn.feature_selection.SelectPercentile"><code class="xref py py-class docutils literal notranslate"><span class="pre">SelectPercentile</span></code></a> removes all but a user-specified highest scoring
percentage of features</p></li>
<li><p>using common univariate statistical tests for each feature:
false positive rate <a class="reference internal" href="generated/sklearn.feature_selection.SelectFpr.html#sklearn.feature_selection.SelectFpr" title="sklearn.feature_selection.SelectFpr"><code class="xref py py-class docutils literal notranslate"><span class="pre">SelectFpr</span></code></a>, false discovery rate
<a class="reference internal" href="generated/sklearn.feature_selection.SelectFdr.html#sklearn.feature_selection.SelectFdr" title="sklearn.feature_selection.SelectFdr"><code class="xref py py-class docutils literal notranslate"><span class="pre">SelectFdr</span></code></a>, or family wise error <a class="reference internal" href="generated/sklearn.feature_selection.SelectFwe.html#sklearn.feature_selection.SelectFwe" title="sklearn.feature_selection.SelectFwe"><code class="xref py py-class docutils literal notranslate"><span class="pre">SelectFwe</span></code></a>.</p></li>
<li><p><a class="reference internal" href="generated/sklearn.feature_selection.GenericUnivariateSelect.html#sklearn.feature_selection.GenericUnivariateSelect" title="sklearn.feature_selection.GenericUnivariateSelect"><code class="xref py py-class docutils literal notranslate"><span class="pre">GenericUnivariateSelect</span></code></a> allows to perform univariate feature
selection with a configurable strategy. This allows to select the best
univariate selection strategy with hyper-parameter search estimator.</p></li>
</ul>
<p>For instance, we can use a F-test to retrieve the two
best features for a dataset as follows:</p>
<div class="doctest highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="kn">from</span><span class="w"> </span><span class="nn">sklearn.datasets</span><span class="w"> </span><span class="kn">import</span> <span class="n">load_iris</span>
<span class="gp">>>> </span><span class="kn">from</span><span class="w"> </span><span class="nn">sklearn.feature_selection</span><span class="w"> </span><span class="kn">import</span> <span class="n">SelectKBest</span>
<span class="gp">>>> </span><span class="kn">from</span><span class="w"> </span><span class="nn">sklearn.feature_selection</span><span class="w"> </span><span class="kn">import</span> <span class="n">f_classif</span>
<span class="gp">>>> </span><span class="n">X</span><span class="p">,</span> <span class="n">y</span> <span class="o">=</span> <span class="n">load_iris</span><span class="p">(</span><span class="n">return_X_y</span><span class="o">=</span><span class="kc">True</span><span class="p">)</span>
<span class="gp">>>> </span><span class="n">X</span><span class="o">.</span><span class="n">shape</span>
<span class="go">(150, 4)</span>
<span class="gp">>>> </span><span class="n">X_new</span> <span class="o">=</span> <span class="n">SelectKBest</span><span class="p">(</span><span class="n">f_classif</span><span class="p">,</span> <span class="n">k</span><span class="o">=</span><span class="mi">2</span><span class="p">)</span><span class="o">.</span><span class="n">fit_transform</span><span class="p">(</span><span class="n">X</span><span class="p">,</span> <span class="n">y</span><span class="p">)</span>
<span class="gp">>>> </span><span class="n">X_new</span><span class="o">.</span><span class="n">shape</span>
<span class="go">(150, 2)</span>
</pre></div>
</div>
<p>These objects take as input a scoring function that returns univariate scores
and p-values (or only scores for <a class="reference internal" href="generated/sklearn.feature_selection.SelectKBest.html#sklearn.feature_selection.SelectKBest" title="sklearn.feature_selection.SelectKBest"><code class="xref py py-class docutils literal notranslate"><span class="pre">SelectKBest</span></code></a> and
<a class="reference internal" href="generated/sklearn.feature_selection.SelectPercentile.html#sklearn.feature_selection.SelectPercentile" title="sklearn.feature_selection.SelectPercentile"><code class="xref py py-class docutils literal notranslate"><span class="pre">SelectPercentile</span></code></a>):</p>
<ul class="simple">
<li><p>For regression: <a class="reference internal" href="generated/sklearn.feature_selection.r_regression.html#sklearn.feature_selection.r_regression" title="sklearn.feature_selection.r_regression"><code class="xref py py-func docutils literal notranslate"><span class="pre">r_regression</span></code></a>, <a class="reference internal" href="generated/sklearn.feature_selection.f_regression.html#sklearn.feature_selection.f_regression" title="sklearn.feature_selection.f_regression"><code class="xref py py-func docutils literal notranslate"><span class="pre">f_regression</span></code></a>, <a class="reference internal" href="generated/sklearn.feature_selection.mutual_info_regression.html#sklearn.feature_selection.mutual_info_regression" title="sklearn.feature_selection.mutual_info_regression"><code class="xref py py-func docutils literal notranslate"><span class="pre">mutual_info_regression</span></code></a></p></li>
<li><p>For classification: <a class="reference internal" href="generated/sklearn.feature_selection.chi2.html#sklearn.feature_selection.chi2" title="sklearn.feature_selection.chi2"><code class="xref py py-func docutils literal notranslate"><span class="pre">chi2</span></code></a>, <a class="reference internal" href="generated/sklearn.feature_selection.f_classif.html#sklearn.feature_selection.f_classif" title="sklearn.feature_selection.f_classif"><code class="xref py py-func docutils literal notranslate"><span class="pre">f_classif</span></code></a>, <a class="reference internal" href="generated/sklearn.feature_selection.mutual_info_classif.html#sklearn.feature_selection.mutual_info_classif" title="sklearn.feature_selection.mutual_info_classif"><code class="xref py py-func docutils literal notranslate"><span class="pre">mutual_info_classif</span></code></a></p></li>
</ul>
<p>The methods based on F-test estimate the degree of linear dependency between
two random variables. On the other hand, mutual information methods can capture
any kind of statistical dependency, but being nonparametric, they require more
samples for accurate estimation. Note that the <span class="math notranslate nohighlight">\(\chi^2\)</span>-test should only be
applied to non-negative features, such as frequencies.</p>
<aside class="topic">
<p class="topic-title">Feature selection with sparse data</p>
<p>If you use sparse data (i.e. data represented as sparse matrices),
<a class="reference internal" href="generated/sklearn.feature_selection.chi2.html#sklearn.feature_selection.chi2" title="sklearn.feature_selection.chi2"><code class="xref py py-func docutils literal notranslate"><span class="pre">chi2</span></code></a>, <a class="reference internal" href="generated/sklearn.feature_selection.mutual_info_regression.html#sklearn.feature_selection.mutual_info_regression" title="sklearn.feature_selection.mutual_info_regression"><code class="xref py py-func docutils literal notranslate"><span class="pre">mutual_info_regression</span></code></a>, <a class="reference internal" href="generated/sklearn.feature_selection.mutual_info_classif.html#sklearn.feature_selection.mutual_info_classif" title="sklearn.feature_selection.mutual_info_classif"><code class="xref py py-func docutils literal notranslate"><span class="pre">mutual_info_classif</span></code></a>
will deal with the data without making it dense.</p>
</aside>
<div class="admonition warning">
<p class="admonition-title">Warning</p>
<p>Beware not to use a regression scoring function with a classification
problem, you will get useless results.</p>
</div>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>The <a class="reference internal" href="generated/sklearn.feature_selection.SelectPercentile.html#sklearn.feature_selection.SelectPercentile" title="sklearn.feature_selection.SelectPercentile"><code class="xref py py-class docutils literal notranslate"><span class="pre">SelectPercentile</span></code></a> and <a class="reference internal" href="generated/sklearn.feature_selection.SelectKBest.html#sklearn.feature_selection.SelectKBest" title="sklearn.feature_selection.SelectKBest"><code class="xref py py-class docutils literal notranslate"><span class="pre">SelectKBest</span></code></a> support unsupervised
feature selection as well. One needs to provide a <code class="docutils literal notranslate"><span class="pre">score_func</span></code> where <code class="docutils literal notranslate"><span class="pre">y=None</span></code>.
The <code class="docutils literal notranslate"><span class="pre">score_func</span></code> should use internally <code class="docutils literal notranslate"><span class="pre">X</span></code> to compute the scores.</p>
</div>
<p class="rubric">Examples</p>
<ul class="simple">
<li><p><a class="reference internal" href="../auto_examples/feature_selection/plot_feature_selection.html#sphx-glr-auto-examples-feature-selection-plot-feature-selection-py"><span class="std std-ref">Univariate Feature Selection</span></a></p></li>
<li><p><a class="reference internal" href="../auto_examples/feature_selection/plot_f_test_vs_mi.html#sphx-glr-auto-examples-feature-selection-plot-f-test-vs-mi-py"><span class="std std-ref">Comparison of F-test and mutual information</span></a></p></li>
</ul>
</section>
<section id="recursive-feature-elimination">
<span id="rfe"></span><h2><span class="section-number">1.13.3. </span>Recursive feature elimination<a class="headerlink" href="#recursive-feature-elimination" title="Link to this heading">#</a></h2>
<p>Given an external estimator that assigns weights to features (e.g., the
coefficients of a linear model), the goal of recursive feature elimination (<a class="reference internal" href="generated/sklearn.feature_selection.RFE.html#sklearn.feature_selection.RFE" title="sklearn.feature_selection.RFE"><code class="xref py py-class docutils literal notranslate"><span class="pre">RFE</span></code></a>)
is to select features by recursively considering smaller and smaller sets of
features. First, the estimator is trained on the initial set of features and
the importance of each feature is obtained either through any specific attribute
(such as <code class="docutils literal notranslate"><span class="pre">coef_</span></code>, <code class="docutils literal notranslate"><span class="pre">feature_importances_</span></code>) or callable. Then, the least important
features are pruned from the current set of features. That procedure is recursively
repeated on the pruned set until the desired number of features to select is
eventually reached.</p>
<p><a class="reference internal" href="generated/sklearn.feature_selection.RFECV.html#sklearn.feature_selection.RFECV" title="sklearn.feature_selection.RFECV"><code class="xref py py-class docutils literal notranslate"><span class="pre">RFECV</span></code></a> performs RFE in a cross-validation loop to find the optimal
number of features. In more details, the number of features selected is tuned
automatically by fitting an <a class="reference internal" href="generated/sklearn.feature_selection.RFE.html#sklearn.feature_selection.RFE" title="sklearn.feature_selection.RFE"><code class="xref py py-class docutils literal notranslate"><span class="pre">RFE</span></code></a> selector on the different
cross-validation splits (provided by the <code class="docutils literal notranslate"><span class="pre">cv</span></code> parameter). The performance
of the <a class="reference internal" href="generated/sklearn.feature_selection.RFE.html#sklearn.feature_selection.RFE" title="sklearn.feature_selection.RFE"><code class="xref py py-class docutils literal notranslate"><span class="pre">RFE</span></code></a> selector is evaluated using <code class="docutils literal notranslate"><span class="pre">scorer</span></code> for different numbers
of selected features and aggregated together. Finally, the scores are averaged
across folds and the number of features selected is set to the number of
features that maximize the cross-validation score.</p>
<p class="rubric">Examples</p>
<ul class="simple">
<li><p><a class="reference internal" href="../auto_examples/feature_selection/plot_rfe_digits.html#sphx-glr-auto-examples-feature-selection-plot-rfe-digits-py"><span class="std std-ref">Recursive feature elimination</span></a>: A recursive feature elimination example
showing the relevance of pixels in a digit classification task.</p></li>
<li><p><a class="reference internal" href="../auto_examples/feature_selection/plot_rfe_with_cross_validation.html#sphx-glr-auto-examples-feature-selection-plot-rfe-with-cross-validation-py"><span class="std std-ref">Recursive feature elimination with cross-validation</span></a>: A recursive feature
elimination example with automatic tuning of the number of features
selected with cross-validation.</p></li>
</ul>
</section>
<section id="feature-selection-using-selectfrommodel">
<span id="select-from-model"></span><h2><span class="section-number">1.13.4. </span>Feature selection using SelectFromModel<a class="headerlink" href="#feature-selection-using-selectfrommodel" title="Link to this heading">#</a></h2>
<p><a class="reference internal" href="generated/sklearn.feature_selection.SelectFromModel.html#sklearn.feature_selection.SelectFromModel" title="sklearn.feature_selection.SelectFromModel"><code class="xref py py-class docutils literal notranslate"><span class="pre">SelectFromModel</span></code></a> is a meta-transformer that can be used alongside any
estimator that assigns importance to each feature through a specific attribute (such as
<code class="docutils literal notranslate"><span class="pre">coef_</span></code>, <code class="docutils literal notranslate"><span class="pre">feature_importances_</span></code>) or via an <code class="docutils literal notranslate"><span class="pre">importance_getter</span></code> callable after fitting.
The features are considered unimportant and removed if the corresponding
importance of the feature values is below the provided
<code class="docutils literal notranslate"><span class="pre">threshold</span></code> parameter. Apart from specifying the threshold numerically,
there are built-in heuristics for finding a threshold using a string argument.
Available heuristics are “mean”, “median” and float multiples of these like
“0.1*mean”. In combination with the <code class="docutils literal notranslate"><span class="pre">threshold</span></code> criteria, one can use the
<code class="docutils literal notranslate"><span class="pre">max_features</span></code> parameter to set a limit on the number of features to select.</p>
<p>For examples on how it is to be used refer to the sections below.</p>
<p class="rubric">Examples</p>
<ul class="simple">
<li><p><a class="reference internal" href="../auto_examples/feature_selection/plot_select_from_model_diabetes.html#sphx-glr-auto-examples-feature-selection-plot-select-from-model-diabetes-py"><span class="std std-ref">Model-based and sequential feature selection</span></a></p></li>
</ul>
<section id="l1-based-feature-selection">
<span id="l1-feature-selection"></span><h3><span class="section-number">1.13.4.1. </span>L1-based feature selection<a class="headerlink" href="#l1-based-feature-selection" title="Link to this heading">#</a></h3>
<p><a class="reference internal" href="linear_model.html#linear-model"><span class="std std-ref">Linear models</span></a> penalized with the L1 norm have
sparse solutions: many of their estimated coefficients are zero. When the goal
is to reduce the dimensionality of the data to use with another classifier,
they can be used along with <a class="reference internal" href="generated/sklearn.feature_selection.SelectFromModel.html#sklearn.feature_selection.SelectFromModel" title="sklearn.feature_selection.SelectFromModel"><code class="xref py py-class docutils literal notranslate"><span class="pre">SelectFromModel</span></code></a>
to select the non-zero coefficients. In particular, sparse estimators useful
for this purpose are the <a class="reference internal" href="generated/sklearn.linear_model.Lasso.html#sklearn.linear_model.Lasso" title="sklearn.linear_model.Lasso"><code class="xref py py-class docutils literal notranslate"><span class="pre">Lasso</span></code></a> for regression, and
of <a class="reference internal" href="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">LogisticRegression</span></code></a> and <a class="reference internal" href="generated/sklearn.svm.LinearSVC.html#sklearn.svm.LinearSVC" title="sklearn.svm.LinearSVC"><code class="xref py py-class docutils literal notranslate"><span class="pre">LinearSVC</span></code></a>
for classification:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="kn">from</span><span class="w"> </span><span class="nn">sklearn.svm</span><span class="w"> </span><span class="kn">import</span> <span class="n">LinearSVC</span>
<span class="gp">>>> </span><span class="kn">from</span><span class="w"> </span><span class="nn">sklearn.datasets</span><span class="w"> </span><span class="kn">import</span> <span class="n">load_iris</span>
<span class="gp">>>> </span><span class="kn">from</span><span class="w"> </span><span class="nn">sklearn.feature_selection</span><span class="w"> </span><span class="kn">import</span> <span class="n">SelectFromModel</span>
<span class="gp">>>> </span><span class="n">X</span><span class="p">,</span> <span class="n">y</span> <span class="o">=</span> <span class="n">load_iris</span><span class="p">(</span><span class="n">return_X_y</span><span class="o">=</span><span class="kc">True</span><span class="p">)</span>
<span class="gp">>>> </span><span class="n">X</span><span class="o">.</span><span class="n">shape</span>
<span class="go">(150, 4)</span>
<span class="gp">>>> </span><span class="n">lsvc</span> <span class="o">=</span> <span class="n">LinearSVC</span><span class="p">(</span><span class="n">C</span><span class="o">=</span><span class="mf">0.01</span><span class="p">,</span> <span class="n">penalty</span><span class="o">=</span><span class="s2">"l1"</span><span class="p">,</span> <span class="n">dual</span><span class="o">=</span><span class="kc">False</span><span class="p">)</span><span class="o">.</span><span class="n">fit</span><span class="p">(</span><span class="n">X</span><span class="p">,</span> <span class="n">y</span><span class="p">)</span>
<span class="gp">>>> </span><span class="n">model</span> <span class="o">=</span> <span class="n">SelectFromModel</span><span class="p">(</span><span class="n">lsvc</span><span class="p">,</span> <span class="n">prefit</span><span class="o">=</span><span class="kc">True</span><span class="p">)</span>
<span class="gp">>>> </span><span class="n">X_new</span> <span class="o">=</span> <span class="n">model</span><span class="o">.</span><span class="n">transform</span><span class="p">(</span><span class="n">X</span><span class="p">)</span>
<span class="gp">>>> </span><span class="n">X_new</span><span class="o">.</span><span class="n">shape</span>
<span class="go">(150, 3)</span>
</pre></div>
</div>
<p>With SVMs and logistic regression, the parameter C controls the sparsity:
the smaller C the fewer features selected. With Lasso, the higher the
alpha parameter, the fewer features selected.</p>
<p class="rubric">Examples</p>
<ul class="simple">
<li><p><a class="reference internal" href="../auto_examples/linear_model/plot_lasso_dense_vs_sparse_data.html#sphx-glr-auto-examples-linear-model-plot-lasso-dense-vs-sparse-data-py"><span class="std std-ref">Lasso on dense and sparse data</span></a>.</p></li>
</ul>
<details class="sd-sphinx-override sd-dropdown sd-card sd-mb-3" id="compressive-sensing">
<span id="l1-recovery-and-compressive-sensing"></span><summary class="sd-summary-title sd-card-header">
<span class="sd-summary-text">L1-recovery and compressive sensing<a class="headerlink" href="#l1-recovery-and-compressive-sensing" title="Link to this dropdown">#</a></span><span class="sd-summary-state-marker sd-summary-chevron-right"><svg version="1.1" width="1.5em" height="1.5em" class="sd-octicon sd-octicon-chevron-right" viewBox="0 0 24 24" aria-hidden="true"><path d="M8.72 18.78a.75.75 0 0 1 0-1.06L14.44 12 8.72 6.28a.751.751 0 0 1 .018-1.042.751.751 0 0 1 1.042-.018l6.25 6.25a.75.75 0 0 1 0 1.06l-6.25 6.25a.75.75 0 0 1-1.06 0Z"></path></svg></span></summary><div class="sd-summary-content sd-card-body docutils">
<p class="sd-card-text">For a good choice of alpha, the <a class="reference internal" href="linear_model.html#lasso"><span class="std std-ref">Lasso</span></a> can fully recover the
exact set of non-zero variables using only few observations, provided
certain specific conditions are met. In particular, the number of
samples should be “sufficiently large”, or L1 models will perform at
random, where “sufficiently large” depends on the number of non-zero
coefficients, the logarithm of the number of features, the amount of
noise, the smallest absolute value of non-zero coefficients, and the
structure of the design matrix X. In addition, the design matrix must
display certain specific properties, such as not being too correlated.
On the use of Lasso for sparse signal recovery, see this example on
compressive sensing:
<a class="reference internal" href="../auto_examples/applications/plot_tomography_l1_reconstruction.html#sphx-glr-auto-examples-applications-plot-tomography-l1-reconstruction-py"><span class="std std-ref">Compressive sensing: tomography reconstruction with L1 prior (Lasso)</span></a>.</p>
<p class="sd-card-text">There is no general rule to select an alpha parameter for recovery of
non-zero coefficients. It can be set by cross-validation
(<a class="reference internal" href="generated/sklearn.linear_model.LassoCV.html#sklearn.linear_model.LassoCV" title="sklearn.linear_model.LassoCV"><code class="xref py py-class docutils literal notranslate"><span class="pre">LassoCV</span></code></a> or
<a class="reference internal" href="generated/sklearn.linear_model.LassoLarsCV.html#sklearn.linear_model.LassoLarsCV" title="sklearn.linear_model.LassoLarsCV"><code class="xref py py-class docutils literal notranslate"><span class="pre">LassoLarsCV</span></code></a>), though this may lead to
under-penalized models: including a small number of non-relevant variables
is not detrimental to prediction score. BIC
(<a class="reference internal" href="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">LassoLarsIC</span></code></a>) tends, on the opposite, to set
high values of alpha.</p>
<p class="rubric">References</p>
<p class="sd-card-text">Richard G. Baraniuk “Compressive Sensing”, IEEE Signal
Processing Magazine [120] July 2007
<a class="reference external" href="http://users.isr.ist.utl.pt/~aguiar/CS_notes.pdf">http://users.isr.ist.utl.pt/~aguiar/CS_notes.pdf</a></p>
</div>
</details></section>
<section id="tree-based-feature-selection">
<h3><span class="section-number">1.13.4.2. </span>Tree-based feature selection<a class="headerlink" href="#tree-based-feature-selection" title="Link to this heading">#</a></h3>
<p>Tree-based estimators (see the <a class="reference internal" href="../api/sklearn.tree.html#module-sklearn.tree" title="sklearn.tree"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.tree</span></code></a> module and forest
of trees in the <a class="reference internal" href="../api/sklearn.ensemble.html#module-sklearn.ensemble" title="sklearn.ensemble"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.ensemble</span></code></a> module) can be used to compute
impurity-based feature importances, which in turn can be used to discard irrelevant
features (when coupled with the <a class="reference internal" href="generated/sklearn.feature_selection.SelectFromModel.html#sklearn.feature_selection.SelectFromModel" title="sklearn.feature_selection.SelectFromModel"><code class="xref py py-class docutils literal notranslate"><span class="pre">SelectFromModel</span></code></a>
meta-transformer):</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="kn">from</span><span class="w"> </span><span class="nn">sklearn.ensemble</span><span class="w"> </span><span class="kn">import</span> <span class="n">ExtraTreesClassifier</span>
<span class="gp">>>> </span><span class="kn">from</span><span class="w"> </span><span class="nn">sklearn.datasets</span><span class="w"> </span><span class="kn">import</span> <span class="n">load_iris</span>
<span class="gp">>>> </span><span class="kn">from</span><span class="w"> </span><span class="nn">sklearn.feature_selection</span><span class="w"> </span><span class="kn">import</span> <span class="n">SelectFromModel</span>
<span class="gp">>>> </span><span class="n">X</span><span class="p">,</span> <span class="n">y</span> <span class="o">=</span> <span class="n">load_iris</span><span class="p">(</span><span class="n">return_X_y</span><span class="o">=</span><span class="kc">True</span><span class="p">)</span>
<span class="gp">>>> </span><span class="n">X</span><span class="o">.</span><span class="n">shape</span>
<span class="go">(150, 4)</span>
<span class="gp">>>> </span><span class="n">clf</span> <span class="o">=</span> <span class="n">ExtraTreesClassifier</span><span class="p">(</span><span class="n">n_estimators</span><span class="o">=</span><span class="mi">50</span><span class="p">)</span>
<span class="gp">>>> </span><span class="n">clf</span> <span class="o">=</span> <span class="n">clf</span><span class="o">.</span><span class="n">fit</span><span class="p">(</span><span class="n">X</span><span class="p">,</span> <span class="n">y</span><span class="p">)</span>
<span class="gp">>>> </span><span class="n">clf</span><span class="o">.</span><span class="n">feature_importances_</span>
<span class="go">array([ 0.04..., 0.05..., 0.4..., 0.4...])</span>
<span class="gp">>>> </span><span class="n">model</span> <span class="o">=</span> <span class="n">SelectFromModel</span><span class="p">(</span><span class="n">clf</span><span class="p">,</span> <span class="n">prefit</span><span class="o">=</span><span class="kc">True</span><span class="p">)</span>
<span class="gp">>>> </span><span class="n">X_new</span> <span class="o">=</span> <span class="n">model</span><span class="o">.</span><span class="n">transform</span><span class="p">(</span><span class="n">X</span><span class="p">)</span>
<span class="gp">>>> </span><span class="n">X_new</span><span class="o">.</span><span class="n">shape</span>
<span class="go">(150, 2)</span>
</pre></div>
</div>
<p class="rubric">Examples</p>
<ul class="simple">
<li><p><a class="reference internal" href="../auto_examples/ensemble/plot_forest_importances.html#sphx-glr-auto-examples-ensemble-plot-forest-importances-py"><span class="std std-ref">Feature importances with a forest of trees</span></a>: example on
synthetic data showing the recovery of the actually meaningful features.</p></li>
<li><p><a class="reference internal" href="../auto_examples/inspection/plot_permutation_importance.html#sphx-glr-auto-examples-inspection-plot-permutation-importance-py"><span class="std std-ref">Permutation Importance vs Random Forest Feature Importance (MDI)</span></a>: example
discussing the caveats of using impurity-based feature importances as a proxy for
feature relevance.</p></li>
</ul>
</section>
</section>
<section id="sequential-feature-selection">
<span id="id3"></span><h2><span class="section-number">1.13.5. </span>Sequential Feature Selection<a class="headerlink" href="#sequential-feature-selection" title="Link to this heading">#</a></h2>
<p>Sequential Feature Selection <a class="reference internal" href="#sfs" id="id4"><span>[sfs]</span></a> (SFS) is available in the
<a class="reference internal" href="generated/sklearn.feature_selection.SequentialFeatureSelector.html#sklearn.feature_selection.SequentialFeatureSelector" title="sklearn.feature_selection.SequentialFeatureSelector"><code class="xref py py-class docutils literal notranslate"><span class="pre">SequentialFeatureSelector</span></code></a> transformer.
SFS can be either forward or backward:</p>
<p>Forward-SFS is a greedy procedure that iteratively finds the best new feature
to add to the set of selected features. Concretely, we initially start with
zero features and find the one feature that maximizes a cross-validated score
when an estimator is trained on this single feature. Once that first feature
is selected, we repeat the procedure by adding a new feature to the set of
selected features. The procedure stops when the desired number of selected
features is reached, as determined by the <code class="docutils literal notranslate"><span class="pre">n_features_to_select</span></code> parameter.</p>
<p>Backward-SFS follows the same idea but works in the opposite direction:
instead of starting with no features and greedily adding features, we start
with <em>all</em> the features and greedily <em>remove</em> features from the set. The
<code class="docutils literal notranslate"><span class="pre">direction</span></code> parameter controls whether forward or backward SFS is used.</p>
<details class="sd-sphinx-override sd-dropdown sd-card sd-mb-3" id="details-on-sequential-feature-selection">
<summary class="sd-summary-title sd-card-header">
<span class="sd-summary-text">Details on Sequential Feature Selection<a class="headerlink" href="#details-on-sequential-feature-selection" title="Link to this dropdown">#</a></span><span class="sd-summary-state-marker sd-summary-chevron-right"><svg version="1.1" width="1.5em" height="1.5em" class="sd-octicon sd-octicon-chevron-right" viewBox="0 0 24 24" aria-hidden="true"><path d="M8.72 18.78a.75.75 0 0 1 0-1.06L14.44 12 8.72 6.28a.751.751 0 0 1 .018-1.042.751.751 0 0 1 1.042-.018l6.25 6.25a.75.75 0 0 1 0 1.06l-6.25 6.25a.75.75 0 0 1-1.06 0Z"></path></svg></span></summary><div class="sd-summary-content sd-card-body docutils">
<p class="sd-card-text">In general, forward and backward selection do not yield equivalent results.
Also, one may be much faster than the other depending on the requested number
of selected features: if we have 10 features and ask for 7 selected features,
forward selection would need to perform 7 iterations while backward selection
would only need to perform 3.</p>
<p class="sd-card-text">SFS differs from <a class="reference internal" href="generated/sklearn.feature_selection.RFE.html#sklearn.feature_selection.RFE" title="sklearn.feature_selection.RFE"><code class="xref py py-class docutils literal notranslate"><span class="pre">RFE</span></code></a> and
<a class="reference internal" href="generated/sklearn.feature_selection.SelectFromModel.html#sklearn.feature_selection.SelectFromModel" title="sklearn.feature_selection.SelectFromModel"><code class="xref py py-class docutils literal notranslate"><span class="pre">SelectFromModel</span></code></a> in that it does not
require the underlying model to expose a <code class="docutils literal notranslate"><span class="pre">coef_</span></code> or <code class="docutils literal notranslate"><span class="pre">feature_importances_</span></code>
attribute. It may however be slower considering that more models need to be
evaluated, compared to the other approaches. For example in backward
selection, the iteration going from <code class="docutils literal notranslate"><span class="pre">m</span></code> features to <code class="docutils literal notranslate"><span class="pre">m</span> <span class="pre">-</span> <span class="pre">1</span></code> features using k-fold
cross-validation requires fitting <code class="docutils literal notranslate"><span class="pre">m</span> <span class="pre">*</span> <span class="pre">k</span></code> models, while
<a class="reference internal" href="generated/sklearn.feature_selection.RFE.html#sklearn.feature_selection.RFE" title="sklearn.feature_selection.RFE"><code class="xref py py-class docutils literal notranslate"><span class="pre">RFE</span></code></a> would require only a single fit, and
<a class="reference internal" href="generated/sklearn.feature_selection.SelectFromModel.html#sklearn.feature_selection.SelectFromModel" title="sklearn.feature_selection.SelectFromModel"><code class="xref py py-class docutils literal notranslate"><span class="pre">SelectFromModel</span></code></a> always just does a single
fit and requires no iterations.</p>
<p class="rubric">References</p>
<div role="list" class="citation-list">
<div class="citation" id="sfs" role="doc-biblioentry">
<span class="label"><span class="fn-bracket">[</span><a role="doc-backlink" href="#id4">sfs</a><span class="fn-bracket">]</span></span>
<p class="sd-card-text">Ferri et al, <a class="reference external" href="https://citeseerx.ist.psu.edu/doc_view/pid/5fedabbb3957bbb442802e012d829ee0629a01b6">Comparative study of techniques for
large-scale feature selection</a>.</p>
</div>
</div>
</div>
</details><p class="rubric">Examples</p>
<ul class="simple">
<li><p><a class="reference internal" href="../auto_examples/feature_selection/plot_select_from_model_diabetes.html#sphx-glr-auto-examples-feature-selection-plot-select-from-model-diabetes-py"><span class="std std-ref">Model-based and sequential feature selection</span></a></p></li>
</ul>
</section>
<section id="feature-selection-as-part-of-a-pipeline">
<h2><span class="section-number">1.13.6. </span>Feature selection as part of a pipeline<a class="headerlink" href="#feature-selection-as-part-of-a-pipeline" title="Link to this heading">#</a></h2>
<p>Feature selection is usually used as a pre-processing step before doing
the actual learning. The recommended way to do this in scikit-learn is
to use a <a class="reference internal" href="generated/sklearn.pipeline.Pipeline.html#sklearn.pipeline.Pipeline" title="sklearn.pipeline.Pipeline"><code class="xref py py-class docutils literal notranslate"><span class="pre">Pipeline</span></code></a>:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">clf</span> <span class="o">=</span> <span class="n">Pipeline</span><span class="p">([</span>
<span class="p">(</span><span class="s1">'feature_selection'</span><span class="p">,</span> <span class="n">SelectFromModel</span><span class="p">(</span><span class="n">LinearSVC</span><span class="p">(</span><span class="n">penalty</span><span class="o">=</span><span class="s2">"l1"</span><span class="p">))),</span>
<span class="p">(</span><span class="s1">'classification'</span><span class="p">,</span> <span class="n">RandomForestClassifier</span><span class="p">())</span>
<span class="p">])</span>
<span class="n">clf</span><span class="o">.</span><span class="n">fit</span><span class="p">(</span><span class="n">X</span><span class="p">,</span> <span class="n">y</span><span class="p">)</span>
</pre></div>
</div>
<p>In this snippet we make use of a <a class="reference internal" href="generated/sklearn.svm.LinearSVC.html#sklearn.svm.LinearSVC" title="sklearn.svm.LinearSVC"><code class="xref py py-class docutils literal notranslate"><span class="pre">LinearSVC</span></code></a>
coupled with <a class="reference internal" href="generated/sklearn.feature_selection.SelectFromModel.html#sklearn.feature_selection.SelectFromModel" title="sklearn.feature_selection.SelectFromModel"><code class="xref py py-class docutils literal notranslate"><span class="pre">SelectFromModel</span></code></a>
to evaluate feature importances and select the most relevant features.
Then, a <a class="reference internal" href="generated/sklearn.ensemble.RandomForestClassifier.html#sklearn.ensemble.RandomForestClassifier" title="sklearn.ensemble.RandomForestClassifier"><code class="xref py py-class docutils literal notranslate"><span class="pre">RandomForestClassifier</span></code></a> is trained on the
transformed output, i.e. using only relevant features. You can perform
similar operations with the other feature selection methods and also
classifiers that provide a way to evaluate feature importances of course.
See the <a class="reference internal" href="generated/sklearn.pipeline.Pipeline.html#sklearn.pipeline.Pipeline" title="sklearn.pipeline.Pipeline"><code class="xref py py-class docutils literal notranslate"><span class="pre">Pipeline</span></code></a> examples for more details.</p>
</section>
</section>