forked from scikit-learn/scikit-learn.github.io
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathreal_world.html
1205 lines (1114 loc) · 76.5 KB
/
real_world.html
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
<!DOCTYPE html>
<!--[if IE 8]><html class="no-js lt-ie9" lang="en" > <![endif]-->
<!--[if gt IE 8]><!--> <html class="no-js" lang="en" > <!--<![endif]-->
<head>
<meta charset="utf-8">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<title>7.2. Real world datasets — scikit-learn 1.0.dev0 documentation</title>
<link rel="canonical" href="http://scikit-learn.org/stable/datasets/real_world.html" />
<link rel="shortcut icon" href="../_static/favicon.ico"/>
<link rel="stylesheet" href="../_static/css/vendor/bootstrap.min.css" type="text/css" />
<link rel="stylesheet" href="../_static/gallery.css" type="text/css" />
<link rel="stylesheet" href="../_static/gallery-binder.css" type="text/css" />
<link rel="stylesheet" href="../_static/gallery-dataframe.css" type="text/css" />
<link rel="stylesheet" href="../_static/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/jquery.js"></script>
</head>
<body>
<nav id="navbar" class="sk-docs-navbar navbar navbar-expand-md navbar-light bg-light py-0">
<div class="container-fluid sk-docs-container px-0">
<a class="navbar-brand py-0" href="../index.html">
<img
class="sk-brand-img"
src="../_static/scikit-learn-logo-small.png"
alt="logo"/>
</a>
<button
id="sk-navbar-toggler"
class="navbar-toggler"
type="button"
data-toggle="collapse"
data-target="#navbarSupportedContent"
aria-controls="navbarSupportedContent"
aria-expanded="false"
aria-label="Toggle navigation"
>
<span class="navbar-toggler-icon"></span>
</button>
<div class="sk-navbar-collapse collapse navbar-collapse" id="navbarSupportedContent">
<ul class="navbar-nav mr-auto">
<li class="nav-item">
<a class="sk-nav-link nav-link" href="../install.html">Install</a>
</li>
<li class="nav-item">
<a class="sk-nav-link nav-link" href="../user_guide.html">User Guide</a>
</li>
<li class="nav-item">
<a class="sk-nav-link nav-link" href="../modules/classes.html">API</a>
</li>
<li class="nav-item">
<a class="sk-nav-link nav-link" href="../auto_examples/index.html">Examples</a>
</li>
<li class="nav-item">
<a class="sk-nav-link nav-link nav-more-item-mobile-items" href="../getting_started.html">Getting Started</a>
</li>
<li class="nav-item">
<a class="sk-nav-link nav-link nav-more-item-mobile-items" href="../tutorial/index.html">Tutorial</a>
</li>
<li class="nav-item">
<a class="sk-nav-link nav-link nav-more-item-mobile-items" href="../whats_new/v1.0.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="../developers/index.html">Development</a>
</li>
<li class="nav-item">
<a class="sk-nav-link nav-link nav-more-item-mobile-items" href="../faq.html">FAQ</a>
</li>
<li class="nav-item">
<a class="sk-nav-link nav-link nav-more-item-mobile-items" href="../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="../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="../whats_new/v1.0.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="../developers/index.html">Development</a>
<a class="sk-nav-dropdown-item dropdown-item" href="../faq.html">FAQ</a>
<a class="sk-nav-dropdown-item dropdown-item" href="../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="../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="sk-sidebar-toc-logo">
<a href="../index.html">
<img
class="sk-brand-img"
src="../_static/scikit-learn-logo-small.png"
alt="logo"/>
</a>
</div>
<div class="btn-group w-100 mb-2" role="group" aria-label="rellinks">
<a href="toy_dataset.html" role="button" class="btn sk-btn-rellink py-1" sk-rellink-tooltip="7.1. Toy datasets">Prev</a><a href="../datasets.html" role="button" class="btn sk-btn-rellink py-1" sk-rellink-tooltip="7. Dataset loading utilities">Up</a>
<a href="sample_generators.html" role="button" class="btn sk-btn-rellink py-1" sk-rellink-tooltip="7.3. Generated datasets">Next</a>
</div>
<div class="alert alert-danger p-1 mb-2" role="alert">
<p class="text-center mb-0">
<strong>scikit-learn 1.0.dev0</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 href="../user_guide.html" class="sk-toc-active">User Guide</a>
</li>
<ul>
<li>
<a href="../supervised_learning.html" class="">1. Supervised learning</a>
</li>
<li>
<a href="../unsupervised_learning.html" class="">2. Unsupervised learning</a>
</li>
<li>
<a href="../model_selection.html" class="">3. Model selection and evaluation</a>
</li>
<li>
<a href="../inspection.html" class="">4. Inspection</a>
</li>
<li>
<a href="../visualizations.html" class="">5. Visualizations</a>
</li>
<li>
<a href="../data_transforms.html" class="">6. Dataset transformations</a>
</li>
<li>
<a href="../datasets.html" class="sk-toc-active">7. Dataset loading utilities</a>
<ul>
<li class="sk-toctree-l3">
<a href="toy_dataset.html">7.1. Toy datasets</a>
</li>
<li class="sk-toctree-l3">
<a href="">7.2. Real world datasets</a>
</li>
<li class="sk-toctree-l3">
<a href="sample_generators.html">7.3. Generated datasets</a>
</li>
<li class="sk-toctree-l3">
<a href="loading_other_datasets.html">7.4. Loading other datasets</a>
</li>
</ul>
</li>
<li>
<a href="../computing.html" class="">8. Computing with scikit-learn</a>
</li>
<li>
<a href="../modules/model_persistence.html" class="">9. Model persistence</a>
</li>
<li>
<a href="../common_pitfalls.html" class="">10. Common pitfalls and recommended practices</a>
</li>
</ul>
</ul>
</div>
</div>
</div>
<div id="sk-page-content-wrapper">
<div class="sk-page-content container-fluid body px-md-3" role="main">
<div class="section" id="real-world-datasets">
<span id="id1"></span><h1><span class="section-number">7.2. </span>Real world datasets<a class="headerlink" href="#real-world-datasets" title="Permalink to this headline">¶</a></h1>
<p>scikit-learn provides tools to load larger datasets, downloading them if
necessary.</p>
<p>They can be loaded using the following functions:</p>
<table class="longtable docutils align-default">
<colgroup>
<col style="width: 10%" />
<col style="width: 90%" />
</colgroup>
<tbody>
<tr class="row-odd"><td><p><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-obj docutils literal notranslate"><span class="pre">fetch_olivetti_faces</span></code></a>(*[, data_home, …])</p></td>
<td><p>Load the Olivetti faces data-set from AT&T (classification).</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="../modules/generated/sklearn.datasets.fetch_20newsgroups.html#sklearn.datasets.fetch_20newsgroups" title="sklearn.datasets.fetch_20newsgroups"><code class="xref py py-obj docutils literal notranslate"><span class="pre">fetch_20newsgroups</span></code></a>(*[, data_home, subset, …])</p></td>
<td><p>Load the filenames and data from the 20 newsgroups dataset (classification).</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="../modules/generated/sklearn.datasets.fetch_20newsgroups_vectorized.html#sklearn.datasets.fetch_20newsgroups_vectorized" title="sklearn.datasets.fetch_20newsgroups_vectorized"><code class="xref py py-obj docutils literal notranslate"><span class="pre">fetch_20newsgroups_vectorized</span></code></a>(*[, subset, …])</p></td>
<td><p>Load and vectorize the 20 newsgroups dataset (classification).</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="../modules/generated/sklearn.datasets.fetch_lfw_people.html#sklearn.datasets.fetch_lfw_people" title="sklearn.datasets.fetch_lfw_people"><code class="xref py py-obj docutils literal notranslate"><span class="pre">fetch_lfw_people</span></code></a>(*[, data_home, funneled, …])</p></td>
<td><p>Load the Labeled Faces in the Wild (LFW) people dataset (classification).</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="../modules/generated/sklearn.datasets.fetch_lfw_pairs.html#sklearn.datasets.fetch_lfw_pairs" title="sklearn.datasets.fetch_lfw_pairs"><code class="xref py py-obj docutils literal notranslate"><span class="pre">fetch_lfw_pairs</span></code></a>(*[, subset, data_home, …])</p></td>
<td><p>Load the Labeled Faces in the Wild (LFW) pairs dataset (classification).</p></td>
</tr>
<tr class="row-even"><td><p><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-obj docutils literal notranslate"><span class="pre">fetch_covtype</span></code></a>(*[, data_home, …])</p></td>
<td><p>Load the covertype dataset (classification).</p></td>
</tr>
<tr class="row-odd"><td><p><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-obj docutils literal notranslate"><span class="pre">fetch_rcv1</span></code></a>(*[, data_home, subset, …])</p></td>
<td><p>Load the RCV1 multilabel dataset (classification).</p></td>
</tr>
<tr class="row-even"><td><p><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-obj docutils literal notranslate"><span class="pre">fetch_kddcup99</span></code></a>(*[, subset, data_home, …])</p></td>
<td><p>Load the kddcup99 dataset (classification).</p></td>
</tr>
<tr class="row-odd"><td><p><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-obj docutils literal notranslate"><span class="pre">fetch_california_housing</span></code></a>(*[, data_home, …])</p></td>
<td><p>Load the California housing dataset (regression).</p></td>
</tr>
</tbody>
</table>
<div class="section" id="the-olivetti-faces-dataset">
<span id="olivetti-faces-dataset"></span><h2><span class="section-number">7.2.1. </span>The Olivetti faces dataset<a class="headerlink" href="#the-olivetti-faces-dataset" title="Permalink to this headline">¶</a></h2>
<p><a class="reference external" href="http://www.cl.cam.ac.uk/research/dtg/attarchive/facedatabase.html">This dataset contains a set of face images</a> taken between April 1992 and
April 1994 at AT&T Laboratories Cambridge. The
<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">sklearn.datasets.fetch_olivetti_faces</span></code></a> function is the data
fetching / caching function that downloads the data
archive from AT&T.</p>
<p>As described on the original website:</p>
<blockquote>
<div><p>There are ten different images of each of 40 distinct subjects. For some
subjects, the images were taken at different times, varying the lighting,
facial expressions (open / closed eyes, smiling / not smiling) and facial
details (glasses / no glasses). All the images were taken against a dark
homogeneous background with the subjects in an upright, frontal position
(with tolerance for some side movement).</p>
</div></blockquote>
<p><strong>Data Set Characteristics:</strong></p>
<blockquote>
<div><table class="docutils align-default">
<colgroup>
<col style="width: 45%" />
<col style="width: 55%" />
</colgroup>
<tbody>
<tr class="row-odd"><td><p>Classes</p></td>
<td><p>40</p></td>
</tr>
<tr class="row-even"><td><p>Samples total</p></td>
<td><p>400</p></td>
</tr>
<tr class="row-odd"><td><p>Dimensionality</p></td>
<td><p>4096</p></td>
</tr>
<tr class="row-even"><td><p>Features</p></td>
<td><p>real, between 0 and 1</p></td>
</tr>
</tbody>
</table>
</div></blockquote>
<p>The image is quantized to 256 grey levels and stored as unsigned 8-bit
integers; the loader will convert these to floating point values on the
interval [0, 1], which are easier to work with for many algorithms.</p>
<p>The “target” for this database is an integer from 0 to 39 indicating the
identity of the person pictured; however, with only 10 examples per class, this
relatively small dataset is more interesting from an unsupervised or
semi-supervised perspective.</p>
<p>The original dataset consisted of 92 x 112, while the version available here
consists of 64x64 images.</p>
<p>When using these images, please give credit to AT&T Laboratories Cambridge.</p>
</div>
<div class="section" id="the-20-newsgroups-text-dataset">
<span id="newsgroups-dataset"></span><h2><span class="section-number">7.2.2. </span>The 20 newsgroups text dataset<a class="headerlink" href="#the-20-newsgroups-text-dataset" title="Permalink to this headline">¶</a></h2>
<p>The 20 newsgroups dataset comprises around 18000 newsgroups posts on
20 topics split in two subsets: one for training (or development)
and the other one for testing (or for performance evaluation). The split
between the train and test set is based upon a messages posted before
and after a specific date.</p>
<p>This module contains two loaders. The first one,
<a class="reference internal" href="../modules/generated/sklearn.datasets.fetch_20newsgroups.html#sklearn.datasets.fetch_20newsgroups" title="sklearn.datasets.fetch_20newsgroups"><code class="xref py py-func docutils literal notranslate"><span class="pre">sklearn.datasets.fetch_20newsgroups</span></code></a>,
returns a list of the raw texts that can be fed to text feature
extractors such as <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">CountVectorizer</span></code></a>
with custom parameters so as to extract feature vectors.
The second one, <a class="reference internal" href="../modules/generated/sklearn.datasets.fetch_20newsgroups_vectorized.html#sklearn.datasets.fetch_20newsgroups_vectorized" title="sklearn.datasets.fetch_20newsgroups_vectorized"><code class="xref py py-func docutils literal notranslate"><span class="pre">sklearn.datasets.fetch_20newsgroups_vectorized</span></code></a>,
returns ready-to-use features, i.e., it is not necessary to use a feature
extractor.</p>
<p><strong>Data Set Characteristics:</strong></p>
<blockquote>
<div><table class="docutils align-default">
<colgroup>
<col style="width: 63%" />
<col style="width: 37%" />
</colgroup>
<tbody>
<tr class="row-odd"><td><p>Classes</p></td>
<td><p>20</p></td>
</tr>
<tr class="row-even"><td><p>Samples total</p></td>
<td><p>18846</p></td>
</tr>
<tr class="row-odd"><td><p>Dimensionality</p></td>
<td><p>1</p></td>
</tr>
<tr class="row-even"><td><p>Features</p></td>
<td><p>text</p></td>
</tr>
</tbody>
</table>
</div></blockquote>
<div class="section" id="usage">
<h3><span class="section-number">7.2.2.1. </span>Usage<a class="headerlink" href="#usage" title="Permalink to this headline">¶</a></h3>
<p>The <a class="reference internal" href="../modules/generated/sklearn.datasets.fetch_20newsgroups.html#sklearn.datasets.fetch_20newsgroups" title="sklearn.datasets.fetch_20newsgroups"><code class="xref py py-func docutils literal notranslate"><span class="pre">sklearn.datasets.fetch_20newsgroups</span></code></a> function is a data
fetching / caching functions that downloads the data archive from
the original <a class="reference external" href="http://people.csail.mit.edu/jrennie/20Newsgroups/">20 newsgroups website</a>, extracts the archive contents
in the <code class="docutils literal notranslate"><span class="pre">~/scikit_learn_data/20news_home</span></code> folder and calls the
<a class="reference internal" href="../modules/generated/sklearn.datasets.load_files.html#sklearn.datasets.load_files" title="sklearn.datasets.load_files"><code class="xref py py-func docutils literal notranslate"><span class="pre">sklearn.datasets.load_files</span></code></a> on either the training or
testing set folder, or both of them:</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.datasets</span> <span class="kn">import</span> <span class="n">fetch_20newsgroups</span>
<span class="gp">>>> </span><span class="n">newsgroups_train</span> <span class="o">=</span> <span class="n">fetch_20newsgroups</span><span class="p">(</span><span class="n">subset</span><span class="o">=</span><span class="s1">'train'</span><span class="p">)</span>
<span class="gp">>>> </span><span class="kn">from</span> <span class="nn">pprint</span> <span class="kn">import</span> <span class="n">pprint</span>
<span class="gp">>>> </span><span class="n">pprint</span><span class="p">(</span><span class="nb">list</span><span class="p">(</span><span class="n">newsgroups_train</span><span class="o">.</span><span class="n">target_names</span><span class="p">))</span>
<span class="go">['alt.atheism',</span>
<span class="go"> 'comp.graphics',</span>
<span class="go"> 'comp.os.ms-windows.misc',</span>
<span class="go"> 'comp.sys.ibm.pc.hardware',</span>
<span class="go"> 'comp.sys.mac.hardware',</span>
<span class="go"> 'comp.windows.x',</span>
<span class="go"> 'misc.forsale',</span>
<span class="go"> 'rec.autos',</span>
<span class="go"> 'rec.motorcycles',</span>
<span class="go"> 'rec.sport.baseball',</span>
<span class="go"> 'rec.sport.hockey',</span>
<span class="go"> 'sci.crypt',</span>
<span class="go"> 'sci.electronics',</span>
<span class="go"> 'sci.med',</span>
<span class="go"> 'sci.space',</span>
<span class="go"> 'soc.religion.christian',</span>
<span class="go"> 'talk.politics.guns',</span>
<span class="go"> 'talk.politics.mideast',</span>
<span class="go"> 'talk.politics.misc',</span>
<span class="go"> 'talk.religion.misc']</span>
</pre></div>
</div>
<p>The real data lies in the <code class="docutils literal notranslate"><span class="pre">filenames</span></code> and <code class="docutils literal notranslate"><span class="pre">target</span></code> attributes. The target
attribute is the integer index of the category:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="n">newsgroups_train</span><span class="o">.</span><span class="n">filenames</span><span class="o">.</span><span class="n">shape</span>
<span class="go">(11314,)</span>
<span class="gp">>>> </span><span class="n">newsgroups_train</span><span class="o">.</span><span class="n">target</span><span class="o">.</span><span class="n">shape</span>
<span class="go">(11314,)</span>
<span class="gp">>>> </span><span class="n">newsgroups_train</span><span class="o">.</span><span class="n">target</span><span class="p">[:</span><span class="mi">10</span><span class="p">]</span>
<span class="go">array([ 7, 4, 4, 1, 14, 16, 13, 3, 2, 4])</span>
</pre></div>
</div>
<p>It is possible to load only a sub-selection of the categories by passing the
list of the categories to load to the
<a class="reference internal" href="../modules/generated/sklearn.datasets.fetch_20newsgroups.html#sklearn.datasets.fetch_20newsgroups" title="sklearn.datasets.fetch_20newsgroups"><code class="xref py py-func docutils literal notranslate"><span class="pre">sklearn.datasets.fetch_20newsgroups</span></code></a> function:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="n">cats</span> <span class="o">=</span> <span class="p">[</span><span class="s1">'alt.atheism'</span><span class="p">,</span> <span class="s1">'sci.space'</span><span class="p">]</span>
<span class="gp">>>> </span><span class="n">newsgroups_train</span> <span class="o">=</span> <span class="n">fetch_20newsgroups</span><span class="p">(</span><span class="n">subset</span><span class="o">=</span><span class="s1">'train'</span><span class="p">,</span> <span class="n">categories</span><span class="o">=</span><span class="n">cats</span><span class="p">)</span>
<span class="gp">>>> </span><span class="nb">list</span><span class="p">(</span><span class="n">newsgroups_train</span><span class="o">.</span><span class="n">target_names</span><span class="p">)</span>
<span class="go">['alt.atheism', 'sci.space']</span>
<span class="gp">>>> </span><span class="n">newsgroups_train</span><span class="o">.</span><span class="n">filenames</span><span class="o">.</span><span class="n">shape</span>
<span class="go">(1073,)</span>
<span class="gp">>>> </span><span class="n">newsgroups_train</span><span class="o">.</span><span class="n">target</span><span class="o">.</span><span class="n">shape</span>
<span class="go">(1073,)</span>
<span class="gp">>>> </span><span class="n">newsgroups_train</span><span class="o">.</span><span class="n">target</span><span class="p">[:</span><span class="mi">10</span><span class="p">]</span>
<span class="go">array([0, 1, 1, 1, 0, 1, 1, 0, 0, 0])</span>
</pre></div>
</div>
</div>
<div class="section" id="converting-text-to-vectors">
<h3><span class="section-number">7.2.2.2. </span>Converting text to vectors<a class="headerlink" href="#converting-text-to-vectors" title="Permalink to this headline">¶</a></h3>
<p>In order to feed predictive or clustering models with the text data,
one first need to turn the text into vectors of numerical values suitable
for statistical analysis. This can be achieved with the utilities of the
<code class="docutils literal notranslate"><span class="pre">sklearn.feature_extraction.text</span></code> as demonstrated in the following
example that extract <a class="reference external" href="https://en.wikipedia.org/wiki/Tf-idf">TF-IDF</a> vectors of unigram tokens
from a subset of 20news:</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.feature_extraction.text</span> <span class="kn">import</span> <span class="n">TfidfVectorizer</span>
<span class="gp">>>> </span><span class="n">categories</span> <span class="o">=</span> <span class="p">[</span><span class="s1">'alt.atheism'</span><span class="p">,</span> <span class="s1">'talk.religion.misc'</span><span class="p">,</span>
<span class="gp">... </span> <span class="s1">'comp.graphics'</span><span class="p">,</span> <span class="s1">'sci.space'</span><span class="p">]</span>
<span class="gp">>>> </span><span class="n">newsgroups_train</span> <span class="o">=</span> <span class="n">fetch_20newsgroups</span><span class="p">(</span><span class="n">subset</span><span class="o">=</span><span class="s1">'train'</span><span class="p">,</span>
<span class="gp">... </span> <span class="n">categories</span><span class="o">=</span><span class="n">categories</span><span class="p">)</span>
<span class="gp">>>> </span><span class="n">vectorizer</span> <span class="o">=</span> <span class="n">TfidfVectorizer</span><span class="p">()</span>
<span class="gp">>>> </span><span class="n">vectors</span> <span class="o">=</span> <span class="n">vectorizer</span><span class="o">.</span><span class="n">fit_transform</span><span class="p">(</span><span class="n">newsgroups_train</span><span class="o">.</span><span class="n">data</span><span class="p">)</span>
<span class="gp">>>> </span><span class="n">vectors</span><span class="o">.</span><span class="n">shape</span>
<span class="go">(2034, 34118)</span>
</pre></div>
</div>
<p>The extracted TF-IDF vectors are very sparse, with an average of 159 non-zero
components by sample in a more than 30000-dimensional space
(less than .5% non-zero features):</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="n">vectors</span><span class="o">.</span><span class="n">nnz</span> <span class="o">/</span> <span class="nb">float</span><span class="p">(</span><span class="n">vectors</span><span class="o">.</span><span class="n">shape</span><span class="p">[</span><span class="mi">0</span><span class="p">])</span>
<span class="go">159.01327...</span>
</pre></div>
</div>
<p><a class="reference internal" href="../modules/generated/sklearn.datasets.fetch_20newsgroups_vectorized.html#sklearn.datasets.fetch_20newsgroups_vectorized" title="sklearn.datasets.fetch_20newsgroups_vectorized"><code class="xref py py-func docutils literal notranslate"><span class="pre">sklearn.datasets.fetch_20newsgroups_vectorized</span></code></a> is a function which
returns ready-to-use token counts features instead of file names.</p>
</div>
<div class="section" id="filtering-text-for-more-realistic-training">
<h3><span class="section-number">7.2.2.3. </span>Filtering text for more realistic training<a class="headerlink" href="#filtering-text-for-more-realistic-training" title="Permalink to this headline">¶</a></h3>
<p>It is easy for a classifier to overfit on particular things that appear in the
20 Newsgroups data, such as newsgroup headers. Many classifiers achieve very
high F-scores, but their results would not generalize to other documents that
aren’t from this window of time.</p>
<p>For example, let’s look at the results of a multinomial Naive Bayes classifier,
which is fast to train and achieves a decent F-score:</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.naive_bayes</span> <span class="kn">import</span> <span class="n">MultinomialNB</span>
<span class="gp">>>> </span><span class="kn">from</span> <span class="nn">sklearn</span> <span class="kn">import</span> <span class="n">metrics</span>
<span class="gp">>>> </span><span class="n">newsgroups_test</span> <span class="o">=</span> <span class="n">fetch_20newsgroups</span><span class="p">(</span><span class="n">subset</span><span class="o">=</span><span class="s1">'test'</span><span class="p">,</span>
<span class="gp">... </span> <span class="n">categories</span><span class="o">=</span><span class="n">categories</span><span class="p">)</span>
<span class="gp">>>> </span><span class="n">vectors_test</span> <span class="o">=</span> <span class="n">vectorizer</span><span class="o">.</span><span class="n">transform</span><span class="p">(</span><span class="n">newsgroups_test</span><span class="o">.</span><span class="n">data</span><span class="p">)</span>
<span class="gp">>>> </span><span class="n">clf</span> <span class="o">=</span> <span class="n">MultinomialNB</span><span class="p">(</span><span class="n">alpha</span><span class="o">=.</span><span class="mi">01</span><span class="p">)</span>
<span class="gp">>>> </span><span class="n">clf</span><span class="o">.</span><span class="n">fit</span><span class="p">(</span><span class="n">vectors</span><span class="p">,</span> <span class="n">newsgroups_train</span><span class="o">.</span><span class="n">target</span><span class="p">)</span>
<span class="go">MultinomialNB(alpha=0.01, class_prior=None, fit_prior=True)</span>
<span class="gp">>>> </span><span class="n">pred</span> <span class="o">=</span> <span class="n">clf</span><span class="o">.</span><span class="n">predict</span><span class="p">(</span><span class="n">vectors_test</span><span class="p">)</span>
<span class="gp">>>> </span><span class="n">metrics</span><span class="o">.</span><span class="n">f1_score</span><span class="p">(</span><span class="n">newsgroups_test</span><span class="o">.</span><span class="n">target</span><span class="p">,</span> <span class="n">pred</span><span class="p">,</span> <span class="n">average</span><span class="o">=</span><span class="s1">'macro'</span><span class="p">)</span>
<span class="go">0.88213...</span>
</pre></div>
</div>
<p>(The example <a class="reference internal" href="../auto_examples/text/plot_document_classification_20newsgroups.html#sphx-glr-auto-examples-text-plot-document-classification-20newsgroups-py"><span class="std std-ref">Classification of text documents using sparse features</span></a> shuffles
the training and test data, instead of segmenting by time, and in that case
multinomial Naive Bayes gets a much higher F-score of 0.88. Are you suspicious
yet of what’s going on inside this classifier?)</p>
<p>Let’s take a look at what the most informative features are:</p>
<div class="doctest highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="kn">import</span> <span class="nn">numpy</span> <span class="k">as</span> <span class="nn">np</span>
<span class="gp">>>> </span><span class="k">def</span> <span class="nf">show_top10</span><span class="p">(</span><span class="n">classifier</span><span class="p">,</span> <span class="n">vectorizer</span><span class="p">,</span> <span class="n">categories</span><span class="p">):</span>
<span class="gp">... </span> <span class="n">feature_names</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">asarray</span><span class="p">(</span><span class="n">vectorizer</span><span class="o">.</span><span class="n">get_feature_names</span><span class="p">())</span>
<span class="gp">... </span> <span class="k">for</span> <span class="n">i</span><span class="p">,</span> <span class="n">category</span> <span class="ow">in</span> <span class="nb">enumerate</span><span class="p">(</span><span class="n">categories</span><span class="p">):</span>
<span class="gp">... </span> <span class="n">top10</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">argsort</span><span class="p">(</span><span class="n">classifier</span><span class="o">.</span><span class="n">coef_</span><span class="p">[</span><span class="n">i</span><span class="p">])[</span><span class="o">-</span><span class="mi">10</span><span class="p">:]</span>
<span class="gp">... </span> <span class="nb">print</span><span class="p">(</span><span class="s2">"</span><span class="si">%s</span><span class="s2">: </span><span class="si">%s</span><span class="s2">"</span> <span class="o">%</span> <span class="p">(</span><span class="n">category</span><span class="p">,</span> <span class="s2">" "</span><span class="o">.</span><span class="n">join</span><span class="p">(</span><span class="n">feature_names</span><span class="p">[</span><span class="n">top10</span><span class="p">])))</span>
<span class="gp">...</span>
<span class="gp">>>> </span><span class="n">show_top10</span><span class="p">(</span><span class="n">clf</span><span class="p">,</span> <span class="n">vectorizer</span><span class="p">,</span> <span class="n">newsgroups_train</span><span class="o">.</span><span class="n">target_names</span><span class="p">)</span>
<span class="go">alt.atheism: edu it and in you that is of to the</span>
<span class="go">comp.graphics: edu in graphics it is for and of to the</span>
<span class="go">sci.space: edu it that is in and space to of the</span>
<span class="go">talk.religion.misc: not it you in is that and to of the</span>
</pre></div>
</div>
<p>You can now see many things that these features have overfit to:</p>
<ul class="simple">
<li><p>Almost every group is distinguished by whether headers such as
<code class="docutils literal notranslate"><span class="pre">NNTP-Posting-Host:</span></code> and <code class="docutils literal notranslate"><span class="pre">Distribution:</span></code> appear more or less often.</p></li>
<li><p>Another significant feature involves whether the sender is affiliated with
a university, as indicated either by their headers or their signature.</p></li>
<li><p>The word “article” is a significant feature, based on how often people quote
previous posts like this: “In article [article ID], [name] <[e-mail address]>
wrote:”</p></li>
<li><p>Other features match the names and e-mail addresses of particular people who
were posting at the time.</p></li>
</ul>
<p>With such an abundance of clues that distinguish newsgroups, the classifiers
barely have to identify topics from text at all, and they all perform at the
same high level.</p>
<p>For this reason, the functions that load 20 Newsgroups data provide a
parameter called <strong>remove</strong>, telling it what kinds of information to strip out
of each file. <strong>remove</strong> should be a tuple containing any subset of
<code class="docutils literal notranslate"><span class="pre">('headers',</span> <span class="pre">'footers',</span> <span class="pre">'quotes')</span></code>, telling it to remove headers, signature
blocks, and quotation blocks respectively.</p>
<div class="doctest highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="n">newsgroups_test</span> <span class="o">=</span> <span class="n">fetch_20newsgroups</span><span class="p">(</span><span class="n">subset</span><span class="o">=</span><span class="s1">'test'</span><span class="p">,</span>
<span class="gp">... </span> <span class="n">remove</span><span class="o">=</span><span class="p">(</span><span class="s1">'headers'</span><span class="p">,</span> <span class="s1">'footers'</span><span class="p">,</span> <span class="s1">'quotes'</span><span class="p">),</span>
<span class="gp">... </span> <span class="n">categories</span><span class="o">=</span><span class="n">categories</span><span class="p">)</span>
<span class="gp">>>> </span><span class="n">vectors_test</span> <span class="o">=</span> <span class="n">vectorizer</span><span class="o">.</span><span class="n">transform</span><span class="p">(</span><span class="n">newsgroups_test</span><span class="o">.</span><span class="n">data</span><span class="p">)</span>
<span class="gp">>>> </span><span class="n">pred</span> <span class="o">=</span> <span class="n">clf</span><span class="o">.</span><span class="n">predict</span><span class="p">(</span><span class="n">vectors_test</span><span class="p">)</span>
<span class="gp">>>> </span><span class="n">metrics</span><span class="o">.</span><span class="n">f1_score</span><span class="p">(</span><span class="n">pred</span><span class="p">,</span> <span class="n">newsgroups_test</span><span class="o">.</span><span class="n">target</span><span class="p">,</span> <span class="n">average</span><span class="o">=</span><span class="s1">'macro'</span><span class="p">)</span>
<span class="go">0.77310...</span>
</pre></div>
</div>
<p>This classifier lost over a lot of its F-score, just because we removed
metadata that has little to do with topic classification.
It loses even more if we also strip this metadata from the training data:</p>
<div class="doctest highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="n">newsgroups_train</span> <span class="o">=</span> <span class="n">fetch_20newsgroups</span><span class="p">(</span><span class="n">subset</span><span class="o">=</span><span class="s1">'train'</span><span class="p">,</span>
<span class="gp">... </span> <span class="n">remove</span><span class="o">=</span><span class="p">(</span><span class="s1">'headers'</span><span class="p">,</span> <span class="s1">'footers'</span><span class="p">,</span> <span class="s1">'quotes'</span><span class="p">),</span>
<span class="gp">... </span> <span class="n">categories</span><span class="o">=</span><span class="n">categories</span><span class="p">)</span>
<span class="gp">>>> </span><span class="n">vectors</span> <span class="o">=</span> <span class="n">vectorizer</span><span class="o">.</span><span class="n">fit_transform</span><span class="p">(</span><span class="n">newsgroups_train</span><span class="o">.</span><span class="n">data</span><span class="p">)</span>
<span class="gp">>>> </span><span class="n">clf</span> <span class="o">=</span> <span class="n">MultinomialNB</span><span class="p">(</span><span class="n">alpha</span><span class="o">=.</span><span class="mi">01</span><span class="p">)</span>
<span class="gp">>>> </span><span class="n">clf</span><span class="o">.</span><span class="n">fit</span><span class="p">(</span><span class="n">vectors</span><span class="p">,</span> <span class="n">newsgroups_train</span><span class="o">.</span><span class="n">target</span><span class="p">)</span>
<span class="go">MultinomialNB(alpha=0.01, class_prior=None, fit_prior=True)</span>
</pre></div>
</div>
<div class="doctest highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="n">vectors_test</span> <span class="o">=</span> <span class="n">vectorizer</span><span class="o">.</span><span class="n">transform</span><span class="p">(</span><span class="n">newsgroups_test</span><span class="o">.</span><span class="n">data</span><span class="p">)</span>
<span class="gp">>>> </span><span class="n">pred</span> <span class="o">=</span> <span class="n">clf</span><span class="o">.</span><span class="n">predict</span><span class="p">(</span><span class="n">vectors_test</span><span class="p">)</span>
<span class="gp">>>> </span><span class="n">metrics</span><span class="o">.</span><span class="n">f1_score</span><span class="p">(</span><span class="n">newsgroups_test</span><span class="o">.</span><span class="n">target</span><span class="p">,</span> <span class="n">pred</span><span class="p">,</span> <span class="n">average</span><span class="o">=</span><span class="s1">'macro'</span><span class="p">)</span>
<span class="go">0.76995...</span>
</pre></div>
</div>
<p>Some other classifiers cope better with this harder version of the task. Try
running <a class="reference internal" href="../auto_examples/model_selection/grid_search_text_feature_extraction.html#sphx-glr-auto-examples-model-selection-grid-search-text-feature-extraction-py"><span class="std std-ref">Sample pipeline for text feature extraction and evaluation</span></a> with and without
the <code class="docutils literal notranslate"><span class="pre">--filter</span></code> option to compare the results.</p>
<div class="topic">
<p class="topic-title">Recommendation</p>
<p>When evaluating text classifiers on the 20 Newsgroups data, you
should strip newsgroup-related metadata. In scikit-learn, you can do this by
setting <code class="docutils literal notranslate"><span class="pre">remove=('headers',</span> <span class="pre">'footers',</span> <span class="pre">'quotes')</span></code>. The F-score will be
lower because it is more realistic.</p>
</div>
<div class="topic">
<p class="topic-title">Examples</p>
<ul class="simple">
<li><p><a class="reference internal" href="../auto_examples/model_selection/grid_search_text_feature_extraction.html#sphx-glr-auto-examples-model-selection-grid-search-text-feature-extraction-py"><span class="std std-ref">Sample pipeline for text feature extraction and evaluation</span></a></p></li>
<li><p><a class="reference internal" href="../auto_examples/text/plot_document_classification_20newsgroups.html#sphx-glr-auto-examples-text-plot-document-classification-20newsgroups-py"><span class="std std-ref">Classification of text documents using sparse features</span></a></p></li>
</ul>
</div>
</div>
</div>
<div class="section" id="the-labeled-faces-in-the-wild-face-recognition-dataset">
<span id="labeled-faces-in-the-wild-dataset"></span><h2><span class="section-number">7.2.3. </span>The Labeled Faces in the Wild face recognition dataset<a class="headerlink" href="#the-labeled-faces-in-the-wild-face-recognition-dataset" title="Permalink to this headline">¶</a></h2>
<p>This dataset is a collection of JPEG pictures of famous people collected
over the internet, all details are available on the official website:</p>
<blockquote>
<div><p><a class="reference external" href="http://vis-www.cs.umass.edu/lfw/">http://vis-www.cs.umass.edu/lfw/</a></p>
</div></blockquote>
<p>Each picture is centered on a single face. The typical task is called
Face Verification: given a pair of two pictures, a binary classifier
must predict whether the two images are from the same person.</p>
<p>An alternative task, Face Recognition or Face Identification is:
given the picture of the face of an unknown person, identify the name
of the person by referring to a gallery of previously seen pictures of
identified persons.</p>
<p>Both Face Verification and Face Recognition are tasks that are typically
performed on the output of a model trained to perform Face Detection. The
most popular model for Face Detection is called Viola-Jones and is
implemented in the OpenCV library. The LFW faces were extracted by this
face detector from various online websites.</p>
<p><strong>Data Set Characteristics:</strong></p>
<blockquote>
<div><table class="docutils align-default">
<colgroup>
<col style="width: 43%" />
<col style="width: 58%" />
</colgroup>
<tbody>
<tr class="row-odd"><td><p>Classes</p></td>
<td><p>5749</p></td>
</tr>
<tr class="row-even"><td><p>Samples total</p></td>
<td><p>13233</p></td>
</tr>
<tr class="row-odd"><td><p>Dimensionality</p></td>
<td><p>5828</p></td>
</tr>
<tr class="row-even"><td><p>Features</p></td>
<td><p>real, between 0 and 255</p></td>
</tr>
</tbody>
</table>
</div></blockquote>
<div class="section" id="id2">
<h3><span class="section-number">7.2.3.1. </span>Usage<a class="headerlink" href="#id2" title="Permalink to this headline">¶</a></h3>
<p><code class="docutils literal notranslate"><span class="pre">scikit-learn</span></code> provides two loaders that will automatically download,
cache, parse the metadata files, decode the jpeg and convert the
interesting slices into memmapped numpy arrays. This dataset size is more
than 200 MB. The first load typically takes more than a couple of minutes
to fully decode the relevant part of the JPEG files into numpy arrays. If
the dataset has been loaded once, the following times the loading times
less than 200ms by using a memmapped version memoized on the disk in the
<code class="docutils literal notranslate"><span class="pre">~/scikit_learn_data/lfw_home/</span></code> folder using <code class="docutils literal notranslate"><span class="pre">joblib</span></code>.</p>
<p>The first loader is used for the Face Identification task: a multi-class
classification task (hence supervised learning):</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.datasets</span> <span class="kn">import</span> <span class="n">fetch_lfw_people</span>
<span class="gp">>>> </span><span class="n">lfw_people</span> <span class="o">=</span> <span class="n">fetch_lfw_people</span><span class="p">(</span><span class="n">min_faces_per_person</span><span class="o">=</span><span class="mi">70</span><span class="p">,</span> <span class="n">resize</span><span class="o">=</span><span class="mf">0.4</span><span class="p">)</span>
<span class="gp">>>> </span><span class="k">for</span> <span class="n">name</span> <span class="ow">in</span> <span class="n">lfw_people</span><span class="o">.</span><span class="n">target_names</span><span class="p">:</span>
<span class="gp">... </span> <span class="nb">print</span><span class="p">(</span><span class="n">name</span><span class="p">)</span>
<span class="gp">...</span>
<span class="go">Ariel Sharon</span>
<span class="go">Colin Powell</span>
<span class="go">Donald Rumsfeld</span>
<span class="go">George W Bush</span>
<span class="go">Gerhard Schroeder</span>
<span class="go">Hugo Chavez</span>
<span class="go">Tony Blair</span>
</pre></div>
</div>
<p>The default slice is a rectangular shape around the face, removing
most of the background:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="n">lfw_people</span><span class="o">.</span><span class="n">data</span><span class="o">.</span><span class="n">dtype</span>
<span class="go">dtype('float32')</span>
<span class="gp">>>> </span><span class="n">lfw_people</span><span class="o">.</span><span class="n">data</span><span class="o">.</span><span class="n">shape</span>
<span class="go">(1288, 1850)</span>
<span class="gp">>>> </span><span class="n">lfw_people</span><span class="o">.</span><span class="n">images</span><span class="o">.</span><span class="n">shape</span>
<span class="go">(1288, 50, 37)</span>
</pre></div>
</div>
<p>Each of the <code class="docutils literal notranslate"><span class="pre">1140</span></code> faces is assigned to a single person id in the <code class="docutils literal notranslate"><span class="pre">target</span></code>
array:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="n">lfw_people</span><span class="o">.</span><span class="n">target</span><span class="o">.</span><span class="n">shape</span>
<span class="go">(1288,)</span>
<span class="gp">>>> </span><span class="nb">list</span><span class="p">(</span><span class="n">lfw_people</span><span class="o">.</span><span class="n">target</span><span class="p">[:</span><span class="mi">10</span><span class="p">])</span>
<span class="go">[5, 6, 3, 1, 0, 1, 3, 4, 3, 0]</span>
</pre></div>
</div>
<p>The second loader is typically used for the face verification task: each sample
is a pair of two picture belonging or not to the same person:</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.datasets</span> <span class="kn">import</span> <span class="n">fetch_lfw_pairs</span>
<span class="gp">>>> </span><span class="n">lfw_pairs_train</span> <span class="o">=</span> <span class="n">fetch_lfw_pairs</span><span class="p">(</span><span class="n">subset</span><span class="o">=</span><span class="s1">'train'</span><span class="p">)</span>
<span class="gp">>>> </span><span class="nb">list</span><span class="p">(</span><span class="n">lfw_pairs_train</span><span class="o">.</span><span class="n">target_names</span><span class="p">)</span>
<span class="go">['Different persons', 'Same person']</span>
<span class="gp">>>> </span><span class="n">lfw_pairs_train</span><span class="o">.</span><span class="n">pairs</span><span class="o">.</span><span class="n">shape</span>
<span class="go">(2200, 2, 62, 47)</span>
<span class="gp">>>> </span><span class="n">lfw_pairs_train</span><span class="o">.</span><span class="n">data</span><span class="o">.</span><span class="n">shape</span>
<span class="go">(2200, 5828)</span>
<span class="gp">>>> </span><span class="n">lfw_pairs_train</span><span class="o">.</span><span class="n">target</span><span class="o">.</span><span class="n">shape</span>
<span class="go">(2200,)</span>
</pre></div>
</div>
<p>Both for the <a class="reference internal" href="../modules/generated/sklearn.datasets.fetch_lfw_people.html#sklearn.datasets.fetch_lfw_people" title="sklearn.datasets.fetch_lfw_people"><code class="xref py py-func docutils literal notranslate"><span class="pre">sklearn.datasets.fetch_lfw_people</span></code></a> and
<a class="reference internal" href="../modules/generated/sklearn.datasets.fetch_lfw_pairs.html#sklearn.datasets.fetch_lfw_pairs" title="sklearn.datasets.fetch_lfw_pairs"><code class="xref py py-func docutils literal notranslate"><span class="pre">sklearn.datasets.fetch_lfw_pairs</span></code></a> function it is
possible to get an additional dimension with the RGB color channels by
passing <code class="docutils literal notranslate"><span class="pre">color=True</span></code>, in that case the shape will be
<code class="docutils literal notranslate"><span class="pre">(2200,</span> <span class="pre">2,</span> <span class="pre">62,</span> <span class="pre">47,</span> <span class="pre">3)</span></code>.</p>
<p>The <a class="reference internal" href="../modules/generated/sklearn.datasets.fetch_lfw_pairs.html#sklearn.datasets.fetch_lfw_pairs" title="sklearn.datasets.fetch_lfw_pairs"><code class="xref py py-func docutils literal notranslate"><span class="pre">sklearn.datasets.fetch_lfw_pairs</span></code></a> datasets is subdivided into
3 subsets: the development <code class="docutils literal notranslate"><span class="pre">train</span></code> set, the development <code class="docutils literal notranslate"><span class="pre">test</span></code> set and
an evaluation <code class="docutils literal notranslate"><span class="pre">10_folds</span></code> set meant to compute performance metrics using a
10-folds cross validation scheme.</p>
<div class="topic">
<p class="topic-title">References:</p>
<ul class="simple">
<li><p><a class="reference external" href="http://vis-www.cs.umass.edu/lfw/lfw.pdf">Labeled Faces in the Wild: A Database for Studying Face Recognition
in Unconstrained Environments.</a>
Gary B. Huang, Manu Ramesh, Tamara Berg, and Erik Learned-Miller.
University of Massachusetts, Amherst, Technical Report 07-49, October, 2007.</p></li>
</ul>
</div>
</div>
<div class="section" id="examples">
<h3><span class="section-number">7.2.3.2. </span>Examples<a class="headerlink" href="#examples" title="Permalink to this headline">¶</a></h3>
<p><a class="reference internal" href="../auto_examples/applications/plot_face_recognition.html#sphx-glr-auto-examples-applications-plot-face-recognition-py"><span class="std std-ref">Faces recognition example using eigenfaces and SVMs</span></a></p>
</div>
</div>
<div class="section" id="forest-covertypes">
<span id="covtype-dataset"></span><h2><span class="section-number">7.2.4. </span>Forest covertypes<a class="headerlink" href="#forest-covertypes" title="Permalink to this headline">¶</a></h2>
<p>The samples in this dataset correspond to 30×30m patches of forest in the US,
collected for the task of predicting each patch’s cover type,
i.e. the dominant species of tree.
There are seven covertypes, making this a multiclass classification problem.
Each sample has 54 features, described on the
<a class="reference external" href="https://archive.ics.uci.edu/ml/datasets/Covertype">dataset’s homepage</a>.
Some of the features are boolean indicators,
while others are discrete or continuous measurements.</p>
<p><strong>Data Set Characteristics:</strong></p>
<blockquote>
<div><table class="docutils align-default">
<colgroup>
<col style="width: 59%" />
<col style="width: 41%" />
</colgroup>
<tbody>
<tr class="row-odd"><td><p>Classes</p></td>
<td><p>7</p></td>
</tr>
<tr class="row-even"><td><p>Samples total</p></td>
<td><p>581012</p></td>
</tr>
<tr class="row-odd"><td><p>Dimensionality</p></td>
<td><p>54</p></td>
</tr>
<tr class="row-even"><td><p>Features</p></td>
<td><p>int</p></td>
</tr>
</tbody>
</table>
</div></blockquote>
<p><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">sklearn.datasets.fetch_covtype</span></code></a> will load the covertype dataset;
it returns a dictionary-like ‘Bunch’ object
with the feature matrix in the <code class="docutils literal notranslate"><span class="pre">data</span></code> member
and the target values in <code class="docutils literal notranslate"><span class="pre">target</span></code>. If optional argument ‘as_frame’ is
set to ‘True’, it will return <code class="docutils literal notranslate"><span class="pre">data</span></code> and <code class="docutils literal notranslate"><span class="pre">target</span></code> as pandas
data frame, and there will be an additional member <code class="docutils literal notranslate"><span class="pre">frame</span></code> as well.
The dataset will be downloaded from the web if necessary.</p>
</div>
<div class="section" id="rcv1-dataset">
<span id="id3"></span><h2><span class="section-number">7.2.5. </span>RCV1 dataset<a class="headerlink" href="#rcv1-dataset" title="Permalink to this headline">¶</a></h2>
<p>Reuters Corpus Volume I (RCV1) is an archive of over 800,000 manually
categorized newswire stories made available by Reuters, Ltd. for research
purposes. The dataset is extensively described in <a class="footnote-reference brackets" href="#id6" id="id4">1</a>.</p>
<p><strong>Data Set Characteristics:</strong></p>
<blockquote>
<div><table class="docutils align-default">
<colgroup>
<col style="width: 40%" />
<col style="width: 60%" />
</colgroup>
<tbody>
<tr class="row-odd"><td><p>Classes</p></td>
<td><p>103</p></td>
</tr>
<tr class="row-even"><td><p>Samples total</p></td>
<td><p>804414</p></td>
</tr>
<tr class="row-odd"><td><p>Dimensionality</p></td>
<td><p>47236</p></td>
</tr>
<tr class="row-even"><td><p>Features</p></td>
<td><p>real, between 0 and 1</p></td>
</tr>
</tbody>
</table>
</div></blockquote>
<p><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">sklearn.datasets.fetch_rcv1</span></code></a> will load the following
version: RCV1-v2, vectors, full sets, topics multilabels:</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.datasets</span> <span class="kn">import</span> <span class="n">fetch_rcv1</span>
<span class="gp">>>> </span><span class="n">rcv1</span> <span class="o">=</span> <span class="n">fetch_rcv1</span><span class="p">()</span>
</pre></div>
</div>
<p>It returns a dictionary-like object, with the following attributes:</p>
<p><code class="docutils literal notranslate"><span class="pre">data</span></code>:
The feature matrix is a scipy CSR sparse matrix, with 804414 samples and
47236 features. Non-zero values contains cosine-normalized, log TF-IDF vectors.
A nearly chronological split is proposed in <a class="footnote-reference brackets" href="#id6" id="id5">1</a>: The first 23149 samples are
the training set. The last 781265 samples are the testing set. This follows
the official LYRL2004 chronological split. The array has 0.16% of non zero
values:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="n">rcv1</span><span class="o">.</span><span class="n">data</span><span class="o">.</span><span class="n">shape</span>
<span class="go">(804414, 47236)</span>
</pre></div>
</div>
<p><code class="docutils literal notranslate"><span class="pre">target</span></code>:
The target values are stored in a scipy CSR sparse matrix, with 804414 samples
and 103 categories. Each sample has a value of 1 in its categories, and 0 in
others. The array has 3.15% of non zero values:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="n">rcv1</span><span class="o">.</span><span class="n">target</span><span class="o">.</span><span class="n">shape</span>
<span class="go">(804414, 103)</span>
</pre></div>
</div>
<p><code class="docutils literal notranslate"><span class="pre">sample_id</span></code>:
Each sample can be identified by its ID, ranging (with gaps) from 2286
to 810596:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="n">rcv1</span><span class="o">.</span><span class="n">sample_id</span><span class="p">[:</span><span class="mi">3</span><span class="p">]</span>
<span class="go">array([2286, 2287, 2288], dtype=uint32)</span>
</pre></div>
</div>
<p><code class="docutils literal notranslate"><span class="pre">target_names</span></code>:
The target values are the topics of each sample. Each sample belongs to at
least one topic, and to up to 17 topics. There are 103 topics, each
represented by a string. Their corpus frequencies span five orders of
magnitude, from 5 occurrences for ‘GMIL’, to 381327 for ‘CCAT’:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="n">rcv1</span><span class="o">.</span><span class="n">target_names</span><span class="p">[:</span><span class="mi">3</span><span class="p">]</span><span class="o">.</span><span class="n">tolist</span><span class="p">()</span>
<span class="go">['E11', 'ECAT', 'M11']</span>
</pre></div>
</div>
<p>The dataset will be downloaded from the <a class="reference external" href="http://jmlr.csail.mit.edu/papers/volume5/lewis04a/">rcv1 homepage</a> if necessary.
The compressed size is about 656 MB.</p>
<div class="topic">
<p class="topic-title">References</p>
<dl class="footnote brackets">
<dt class="label" id="id6"><span class="brackets">1</span><span class="fn-backref">(<a href="#id4">1</a>,<a href="#id5">2</a>)</span></dt>
<dd><p>Lewis, D. D., Yang, Y., Rose, T. G., & Li, F. (2004).
RCV1: A new benchmark collection for text categorization research.
The Journal of Machine Learning Research, 5, 361-397.</p>
</dd>
</dl>
</div>
</div>
<div class="section" id="kddcup-99-dataset">
<span id="kddcup99-dataset"></span><h2><span class="section-number">7.2.6. </span>Kddcup 99 dataset<a class="headerlink" href="#kddcup-99-dataset" title="Permalink to this headline">¶</a></h2>
<p>The KDD Cup ‘99 dataset was created by processing the tcpdump portions
of the 1998 DARPA Intrusion Detection System (IDS) Evaluation dataset,
created by MIT Lincoln Lab <a class="footnote-reference brackets" href="#id9" id="id7">2</a>. The artificial data (described on the <a class="reference external" href="https://kdd.ics.uci.edu/databases/kddcup99/kddcup99.html">dataset’s
homepage</a>) was
generated using a closed network and hand-injected attacks to produce a
large number of different types of attack with normal activity in the
background. As the initial goal was to produce a large training set for
supervised learning algorithms, there is a large proportion (80.1%) of
abnormal data which is unrealistic in real world, and inappropriate for
unsupervised anomaly detection which aims at detecting ‘abnormal’ data, i.e.:</p>
<ul class="simple">
<li><p>qualitatively different from normal data</p></li>
<li><p>in large minority among the observations.</p></li>
</ul>
<p>We thus transform the KDD Data set into two different data sets: SA and SF.</p>
<ul class="simple">
<li><p>SA is obtained by simply selecting all the normal data, and a small
proportion of abnormal data to gives an anomaly proportion of 1%.</p></li>
<li><p>SF is obtained as in <a class="footnote-reference brackets" href="#id10" id="id8">3</a>
by simply picking up the data whose attribute logged_in is positive, thus
focusing on the intrusion attack, which gives a proportion of 0.3% of
attack.</p></li>
<li><p>http and smtp are two subsets of SF corresponding with third feature
equal to ‘http’ (resp. to ‘smtp’).</p></li>
</ul>
<p>General KDD structure :</p>
<blockquote>
<div><table class="docutils align-default">
<colgroup>
<col style="width: 28%" />
<col style="width: 72%" />
</colgroup>
<tbody>
<tr class="row-odd"><td><p>Samples total</p></td>
<td><p>4898431</p></td>
</tr>
<tr class="row-even"><td><p>Dimensionality</p></td>
<td><p>41</p></td>
</tr>
<tr class="row-odd"><td><p>Features</p></td>
<td><p>discrete (int) or continuous (float)</p></td>
</tr>
<tr class="row-even"><td><p>Targets</p></td>
<td><p>str, ‘normal.’ or name of the anomaly type</p></td>
</tr>
</tbody>
</table>
<p>SA structure :</p>
<table class="docutils align-default">
<colgroup>
<col style="width: 28%" />
<col style="width: 72%" />
</colgroup>
<tbody>
<tr class="row-odd"><td><p>Samples total</p></td>
<td><p>976158</p></td>
</tr>
<tr class="row-even"><td><p>Dimensionality</p></td>
<td><p>41</p></td>
</tr>
<tr class="row-odd"><td><p>Features</p></td>
<td><p>discrete (int) or continuous (float)</p></td>
</tr>
<tr class="row-even"><td><p>Targets</p></td>
<td><p>str, ‘normal.’ or name of the anomaly type</p></td>
</tr>
</tbody>
</table>
<p>SF structure :</p>
<table class="docutils align-default">
<colgroup>
<col style="width: 28%" />
<col style="width: 72%" />
</colgroup>
<tbody>
<tr class="row-odd"><td><p>Samples total</p></td>
<td><p>699691</p></td>
</tr>
<tr class="row-even"><td><p>Dimensionality</p></td>
<td><p>4</p></td>
</tr>
<tr class="row-odd"><td><p>Features</p></td>
<td><p>discrete (int) or continuous (float)</p></td>
</tr>
<tr class="row-even"><td><p>Targets</p></td>
<td><p>str, ‘normal.’ or name of the anomaly type</p></td>
</tr>
</tbody>
</table>
<p>http structure :</p>
<table class="docutils align-default">
<colgroup>
<col style="width: 28%" />
<col style="width: 72%" />
</colgroup>
<tbody>
<tr class="row-odd"><td><p>Samples total</p></td>
<td><p>619052</p></td>
</tr>
<tr class="row-even"><td><p>Dimensionality</p></td>
<td><p>3</p></td>
</tr>
<tr class="row-odd"><td><p>Features</p></td>
<td><p>discrete (int) or continuous (float)</p></td>
</tr>
<tr class="row-even"><td><p>Targets</p></td>
<td><p>str, ‘normal.’ or name of the anomaly type</p></td>
</tr>
</tbody>
</table>
<p>smtp structure :</p>
<table class="docutils align-default">
<colgroup>
<col style="width: 28%" />
<col style="width: 72%" />
</colgroup>
<tbody>
<tr class="row-odd"><td><p>Samples total</p></td>
<td><p>95373</p></td>
</tr>
<tr class="row-even"><td><p>Dimensionality</p></td>
<td><p>3</p></td>
</tr>
<tr class="row-odd"><td><p>Features</p></td>
<td><p>discrete (int) or continuous (float)</p></td>
</tr>
<tr class="row-even"><td><p>Targets</p></td>
<td><p>str, ‘normal.’ or name of the anomaly type</p></td>
</tr>
</tbody>
</table>
</div></blockquote>
<p><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">sklearn.datasets.fetch_kddcup99</span></code></a> will load the kddcup99 dataset; it
returns a dictionary-like object with the feature matrix in the <code class="docutils literal notranslate"><span class="pre">data</span></code> member
and the target values in <code class="docutils literal notranslate"><span class="pre">target</span></code>. The “as_frame” optional argument converts
<code class="docutils literal notranslate"><span class="pre">data</span></code> into a pandas DataFrame and <code class="docutils literal notranslate"><span class="pre">target</span></code> into a pandas Series. The
dataset will be downloaded from the web if necessary.</p>
<div class="topic">
<p class="topic-title">References</p>
<dl class="footnote brackets">
<dt class="label" id="id9"><span class="brackets"><a class="fn-backref" href="#id7">2</a></span></dt>
<dd><p>Analysis and Results of the 1999 DARPA Off-Line Intrusion
Detection Evaluation, Richard Lippmann, Joshua W. Haines,
David J. Fried, Jonathan Korba, Kumar Das.</p>
</dd>
<dt class="label" id="id10"><span class="brackets"><a class="fn-backref" href="#id8">3</a></span></dt>
<dd><p>K. Yamanishi, J.-I. Takeuchi, G. Williams, and P. Milne. Online
unsupervised outlier detection using finite mixtures with
discounting learning algorithms. In Proceedings of the sixth
ACM SIGKDD international conference on Knowledge discovery
and data mining, pages 320-324. ACM Press, 2000.</p>
</dd>
</dl>
</div>
</div>