-
Notifications
You must be signed in to change notification settings - Fork 81
/
Copy patholder_versions.html
1478 lines (1443 loc) · 131 KB
/
older_versions.html
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
<!DOCTYPE html>
<!--[if IE 8]><html class="no-js lt-ie9" lang="en" > <![endif]-->
<!--[if gt IE 8]><!--> <html class="no-js" lang="en" > <!--<![endif]-->
<head>
<meta charset="utf-8">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<meta name="Description" content="scikit-learn: machine learning in Python">
<title>Version 0.12.1 — scikit-learn 0.22.2 documentation</title>
<link rel="canonical" href="https://scikit-learn.org/stable/whats_new/older_versions.html" />
<link rel="shortcut icon" href="../_static/favicon.ico"/>
<link rel="stylesheet" href="../_static/css/vendor/bootstrap.min.css" type="text/css" />
<link rel="stylesheet" href="../_static/gallery.css" type="text/css" />
<link rel="stylesheet" href="../_static/css/theme.css" type="text/css" />
<script id="documentation_options" data-url_root="../" src="../_static/documentation_options.js"></script>
<script src="../_static/jquery.js"></script>
</head>
<body>
<nav id="navbar" class="sk-docs-navbar navbar navbar-expand-md navbar-light bg-light py-0">
<div class="container-fluid sk-docs-container px-0">
<a class="navbar-brand py-0" href="../index.html">
<img
class="sk-brand-img"
src="../_static/scikit-learn-logo-small.png"
alt="logo"/>
</a>
<button
id="sk-navbar-toggler"
class="navbar-toggler"
type="button"
data-toggle="collapse"
data-target="#navbarSupportedContent"
aria-controls="navbarSupportedContent"
aria-expanded="false"
aria-label="Toggle navigation"
>
<span class="navbar-toggler-icon"></span>
</button>
<div class="sk-navbar-collapse collapse navbar-collapse" id="navbarSupportedContent">
<ul class="navbar-nav mr-auto">
<li class="nav-item">
<a class="sk-nav-link nav-link" href="../install.html">Install</a>
</li>
<li class="nav-item">
<a class="sk-nav-link nav-link" href="../user_guide.html">User Guide</a>
</li>
<li class="nav-item">
<a class="sk-nav-link nav-link" href="../modules/classes.html">API</a>
</li>
<li class="nav-item">
<a class="sk-nav-link nav-link" href="../auto_examples/index.html">Examples</a>
</li>
<li class="nav-item">
<a class="sk-nav-link nav-link nav-more-item-mobile-items" href="../getting_started.html">Getting Started</a>
</li>
<li class="nav-item">
<a class="sk-nav-link nav-link nav-more-item-mobile-items" href="../tutorial/index.html">Tutorial</a>
</li>
<li class="nav-item">
<a class="sk-nav-link nav-link nav-more-item-mobile-items" href="../glossary.html">Glossary</a>
</li>
<li class="nav-item">
<a class="sk-nav-link nav-link nav-more-item-mobile-items" href="../developers/index.html">Development</a>
</li>
<li class="nav-item">
<a class="sk-nav-link nav-link nav-more-item-mobile-items" href="../faq.html">FAQ</a>
</li>
<li class="nav-item">
<a class="sk-nav-link nav-link nav-more-item-mobile-items" href="../related_projects.html">Related packages</a>
</li>
<li class="nav-item">
<a class="sk-nav-link nav-link nav-more-item-mobile-items" href="../roadmap.html">Roadmap</a>
</li>
<li class="nav-item">
<a class="sk-nav-link nav-link nav-more-item-mobile-items" href="../about.html">About us</a>
</li>
<li class="nav-item">
<a class="sk-nav-link nav-link nav-more-item-mobile-items" href="https://github.com/scikit-learn/scikit-learn">GitHub</a>
</li>
<li class="nav-item">
<a class="sk-nav-link nav-link nav-more-item-mobile-items" href="https://scikit-learn.org/dev/versions.html">Other Versions</a>
</li>
<li class="nav-item dropdown nav-more-item-dropdown">
<a class="sk-nav-link nav-link dropdown-toggle" href="#" id="navbarDropdown" role="button" data-toggle="dropdown" aria-haspopup="true" aria-expanded="false">More</a>
<div class="dropdown-menu" aria-labelledby="navbarDropdown">
<a class="sk-nav-dropdown-item dropdown-item" href="../getting_started.html">Getting Started</a>
<a class="sk-nav-dropdown-item dropdown-item" href="../tutorial/index.html">Tutorial</a>
<a class="sk-nav-dropdown-item dropdown-item" href="../glossary.html">Glossary</a>
<a class="sk-nav-dropdown-item dropdown-item" href="../developers/index.html">Development</a>
<a class="sk-nav-dropdown-item dropdown-item" href="../faq.html">FAQ</a>
<a class="sk-nav-dropdown-item dropdown-item" href="../related_projects.html">Related packages</a>
<a class="sk-nav-dropdown-item dropdown-item" href="../roadmap.html">Roadmap</a>
<a class="sk-nav-dropdown-item dropdown-item" href="../about.html">About us</a>
<a class="sk-nav-dropdown-item dropdown-item" href="https://github.com/scikit-learn/scikit-learn">GitHub</a>
<a class="sk-nav-dropdown-item dropdown-item" href="https://scikit-learn.org/dev/versions.html">Other Versions</a>
</div>
</li>
</ul>
<div id="searchbox" role="search">
<div class="searchformwrapper">
<form class="search" action="../search.html" method="get">
<input class="sk-search-text-input" type="text" name="q" aria-labelledby="searchlabel" />
<input class="sk-search-text-btn" type="submit" value="Go" />
</form>
</div>
</div>
</div>
</div>
</nav>
<div class="d-flex" id="sk-doc-wrapper">
<input type="checkbox" name="sk-toggle-checkbox" id="sk-toggle-checkbox">
<label id="sk-sidemenu-toggle" class="sk-btn-toggle-toc btn sk-btn-primary" for="sk-toggle-checkbox">Toggle Menu</label>
<div id="sk-sidebar-wrapper" class="border-right">
<div class="sk-sidebar-toc-wrapper">
<div class="sk-sidebar-toc-logo">
<a href="../index.html">
<img
class="sk-brand-img"
src="../_static/scikit-learn-logo-small.png"
alt="logo"/>
</a>
</div>
<div class="btn-group w-100 mb-2" role="group" aria-label="rellinks">
<a href="v0.13.html" role="button" class="btn sk-btn-rellink py-1" sk-rellink-tooltip="Version 0.13.1">Prev</a><a href="../whats_new.html" role="button" class="btn sk-btn-rellink py-1" sk-rellink-tooltip="Release History">Up</a>
<a href="../roadmap.html" role="button" class="btn sk-btn-rellink py-1" sk-rellink-tooltip="Roadmap">Next</a>
</div>
<div class="alert alert-danger p-1 mb-2" role="alert">
<p class="text-center mb-0">
<strong>scikit-learn 0.22.2</strong><br/>
<a href="http://scikit-learn.org/dev/versions.html">Other versions</a>
</p>
</div>
<div class="alert alert-warning p-1 mb-2" role="alert">
<p class="text-center mb-0">
Please <a class="font-weight-bold" href="../about.html#citing-scikit-learn"><string>cite us</string></a> if you use the software.
</p>
</div>
<div class="sk-sidebar-toc">
<ul>
<li><a class="reference internal" href="#">Version 0.12.1</a><ul>
<li><a class="reference internal" href="#changelog">Changelog</a></li>
<li><a class="reference internal" href="#people">People</a></li>
</ul>
</li>
<li><a class="reference internal" href="#version-0-12">Version 0.12</a><ul>
<li><a class="reference internal" href="#id1">Changelog</a></li>
<li><a class="reference internal" href="#api-changes-summary">API changes summary</a></li>
<li><a class="reference internal" href="#id2">People</a></li>
</ul>
</li>
<li><a class="reference internal" href="#version-0-11">Version 0.11</a><ul>
<li><a class="reference internal" href="#id3">Changelog</a><ul>
<li><a class="reference internal" href="#highlights">Highlights</a></li>
<li><a class="reference internal" href="#other-changes">Other changes</a></li>
</ul>
</li>
<li><a class="reference internal" href="#id4">API changes summary</a></li>
<li><a class="reference internal" href="#id5">People</a></li>
</ul>
</li>
<li><a class="reference internal" href="#version-0-10">Version 0.10</a><ul>
<li><a class="reference internal" href="#id6">Changelog</a></li>
<li><a class="reference internal" href="#id7">API changes summary</a></li>
<li><a class="reference internal" href="#id8">People</a></li>
</ul>
</li>
<li><a class="reference internal" href="#version-0-9">Version 0.9</a><ul>
<li><a class="reference internal" href="#id9">Changelog</a></li>
<li><a class="reference internal" href="#id10">API changes summary</a></li>
<li><a class="reference internal" href="#id11">People</a></li>
</ul>
</li>
<li><a class="reference internal" href="#version-0-8">Version 0.8</a><ul>
<li><a class="reference internal" href="#id12">Changelog</a></li>
<li><a class="reference internal" href="#id13">People</a></li>
</ul>
</li>
<li><a class="reference internal" href="#version-0-7">Version 0.7</a><ul>
<li><a class="reference internal" href="#id14">Changelog</a></li>
<li><a class="reference internal" href="#id15">People</a></li>
</ul>
</li>
<li><a class="reference internal" href="#version-0-6">Version 0.6</a><ul>
<li><a class="reference internal" href="#id16">Changelog</a></li>
<li><a class="reference internal" href="#id17">People</a></li>
</ul>
</li>
<li><a class="reference internal" href="#version-0-5">Version 0.5</a><ul>
<li><a class="reference internal" href="#id18">Changelog</a></li>
<li><a class="reference internal" href="#new-classes">New classes</a></li>
<li><a class="reference internal" href="#documentation">Documentation</a></li>
<li><a class="reference internal" href="#fixes">Fixes</a></li>
<li><a class="reference internal" href="#examples">Examples</a></li>
<li><a class="reference internal" href="#external-dependencies">External dependencies</a></li>
<li><a class="reference internal" href="#removed-modules">Removed modules</a></li>
<li><a class="reference internal" href="#misc">Misc</a></li>
<li><a class="reference internal" href="#authors">Authors</a></li>
</ul>
</li>
<li><a class="reference internal" href="#version-0-4">Version 0.4</a><ul>
<li><a class="reference internal" href="#id19">Changelog</a></li>
<li><a class="reference internal" href="#id20">Authors</a></li>
</ul>
</li>
<li><a class="reference internal" href="#earlier-versions">Earlier versions</a></li>
</ul>
</div>
</div>
</div>
<div id="sk-page-content-wrapper">
<div class="sk-page-content container-fluid body px-md-3" role="main">
<div class="section" id="version-0-12-1">
<span id="changes-0-12-1"></span><h1>Version 0.12.1<a class="headerlink" href="#version-0-12-1" title="Permalink to this headline">¶</a></h1>
<p><strong>October 8, 2012</strong></p>
<p>The 0.12.1 release is a bug-fix release with no additional features, but is
instead a set of bug fixes</p>
<div class="section" id="changelog">
<h2>Changelog<a class="headerlink" href="#changelog" title="Permalink to this headline">¶</a></h2>
<ul class="simple">
<li><p>Improved numerical stability in spectral embedding by <a class="reference external" href="http://gael-varoquaux.info">Gael
Varoquaux</a></p></li>
<li><p>Doctest under windows 64bit by <a class="reference external" href="http://gael-varoquaux.info">Gael Varoquaux</a></p></li>
<li><p>Documentation fixes for elastic net by <a class="reference external" href="https://amueller.github.io/">Andreas Müller</a> and
<a class="reference external" href="http://alexandre.gramfort.net">Alexandre Gramfort</a></p></li>
<li><p>Proper behavior with fortran-ordered NumPy arrays by <a class="reference external" href="http://gael-varoquaux.info">Gael Varoquaux</a></p></li>
<li><p>Make GridSearchCV work with non-CSR sparse matrix by <a class="reference external" href="https://github.com/larsmans">Lars Buitinck</a></p></li>
<li><p>Fix parallel computing in MDS by <a class="reference external" href="http://gael-varoquaux.info">Gael Varoquaux</a></p></li>
<li><p>Fix Unicode support in count vectorizer by <a class="reference external" href="https://amueller.github.io/">Andreas Müller</a></p></li>
<li><p>Fix MinCovDet breaking with X.shape = (3, 1) by <a class="reference external" href="https://github.com/VirgileFritsch">Virgile Fritsch</a></p></li>
<li><p>Fix clone of SGD objects by <a class="reference external" href="https://sites.google.com/site/peterprettenhofer/">Peter Prettenhofer</a></p></li>
<li><p>Stabilize GMM by <a class="reference external" href="https://github.com/VirgileFritsch">Virgile Fritsch</a></p></li>
</ul>
</div>
<div class="section" id="people">
<h2>People<a class="headerlink" href="#people" title="Permalink to this headline">¶</a></h2>
<blockquote>
<div><ul class="simple">
<li><p>14 <a class="reference external" href="https://sites.google.com/site/peterprettenhofer/">Peter Prettenhofer</a></p></li>
<li><p>12 <a class="reference external" href="http://gael-varoquaux.info">Gael Varoquaux</a></p></li>
<li><p>10 <a class="reference external" href="https://amueller.github.io/">Andreas Müller</a></p></li>
<li><p>5 <a class="reference external" href="https://github.com/larsmans">Lars Buitinck</a></p></li>
<li><p>3 <a class="reference external" href="https://github.com/VirgileFritsch">Virgile Fritsch</a></p></li>
<li><p>1 <a class="reference external" href="http://alexandre.gramfort.net">Alexandre Gramfort</a></p></li>
<li><p>1 <a class="reference external" href="http://www.montefiore.ulg.ac.be/~glouppe/">Gilles Louppe</a></p></li>
<li><p>1 <a class="reference external" href="http://www.mblondel.org">Mathieu Blondel</a></p></li>
</ul>
</div></blockquote>
</div>
</div>
<div class="section" id="version-0-12">
<span id="changes-0-12"></span><h1>Version 0.12<a class="headerlink" href="#version-0-12" title="Permalink to this headline">¶</a></h1>
<p><strong>September 4, 2012</strong></p>
<div class="section" id="id1">
<h2>Changelog<a class="headerlink" href="#id1" title="Permalink to this headline">¶</a></h2>
<ul class="simple">
<li><p>Various speed improvements of the <a class="reference internal" href="../modules/tree.html#tree"><span class="std std-ref">decision trees</span></a> module, by
<a class="reference external" href="http://www.montefiore.ulg.ac.be/~glouppe/">Gilles Louppe</a>.</p></li>
<li><p><a class="reference internal" href="../modules/generated/sklearn.ensemble.GradientBoostingRegressor.html#sklearn.ensemble.GradientBoostingRegressor" title="sklearn.ensemble.GradientBoostingRegressor"><code class="xref py py-class docutils literal notranslate"><span class="pre">ensemble.GradientBoostingRegressor</span></code></a> and
<a class="reference internal" href="../modules/generated/sklearn.ensemble.GradientBoostingClassifier.html#sklearn.ensemble.GradientBoostingClassifier" title="sklearn.ensemble.GradientBoostingClassifier"><code class="xref py py-class docutils literal notranslate"><span class="pre">ensemble.GradientBoostingClassifier</span></code></a> now support feature subsampling
via the <code class="docutils literal notranslate"><span class="pre">max_features</span></code> argument, by <a class="reference external" href="https://sites.google.com/site/peterprettenhofer/">Peter Prettenhofer</a>.</p></li>
<li><p>Added Huber and Quantile loss functions to
<a class="reference internal" href="../modules/generated/sklearn.ensemble.GradientBoostingRegressor.html#sklearn.ensemble.GradientBoostingRegressor" title="sklearn.ensemble.GradientBoostingRegressor"><code class="xref py py-class docutils literal notranslate"><span class="pre">ensemble.GradientBoostingRegressor</span></code></a>, by <a class="reference external" href="https://sites.google.com/site/peterprettenhofer/">Peter Prettenhofer</a>.</p></li>
<li><p><a class="reference internal" href="../modules/tree.html#tree"><span class="std std-ref">Decision trees</span></a> and <a class="reference internal" href="../modules/ensemble.html#forest"><span class="std std-ref">forests of randomized trees</span></a>
now support multi-output classification and regression problems, by
<a class="reference external" href="http://www.montefiore.ulg.ac.be/~glouppe/">Gilles Louppe</a>.</p></li>
<li><p>Added <a class="reference internal" href="../modules/generated/sklearn.preprocessing.LabelEncoder.html#sklearn.preprocessing.LabelEncoder" title="sklearn.preprocessing.LabelEncoder"><code class="xref py py-class docutils literal notranslate"><span class="pre">preprocessing.LabelEncoder</span></code></a>, a simple utility class to
normalize labels or transform non-numerical labels, by <a class="reference external" href="http://www.mblondel.org">Mathieu Blondel</a>.</p></li>
<li><p>Added the epsilon-insensitive loss and the ability to make probabilistic
predictions with the modified huber loss in <a class="reference internal" href="../modules/sgd.html#sgd"><span class="std std-ref">Stochastic Gradient Descent</span></a>, by
<a class="reference external" href="http://www.mblondel.org">Mathieu Blondel</a>.</p></li>
<li><p>Added <a class="reference internal" href="../modules/manifold.html#multidimensional-scaling"><span class="std std-ref">Multi-dimensional Scaling (MDS)</span></a>, by Nelle Varoquaux.</p></li>
<li><p>SVMlight file format loader now detects compressed (gzip/bzip2) files and
decompresses them on the fly, by <a class="reference external" href="https://github.com/larsmans">Lars Buitinck</a>.</p></li>
<li><p>SVMlight file format serializer now preserves double precision floating
point values, by <a class="reference external" href="https://twitter.com/ogrisel">Olivier Grisel</a>.</p></li>
<li><p>A common testing framework for all estimators was added, by <a class="reference external" href="https://amueller.github.io/">Andreas Müller</a>.</p></li>
<li><p>Understandable error messages for estimators that do not accept
sparse input by <a class="reference external" href="http://gael-varoquaux.info">Gael Varoquaux</a></p></li>
<li><p>Speedups in hierarchical clustering by <a class="reference external" href="http://gael-varoquaux.info">Gael Varoquaux</a>. In
particular building the tree now supports early stopping. This is
useful when the number of clusters is not small compared to the
number of samples.</p></li>
<li><p>Add MultiTaskLasso and MultiTaskElasticNet for joint feature selection,
by <a class="reference external" href="http://alexandre.gramfort.net">Alexandre Gramfort</a>.</p></li>
<li><p>Added <code class="xref py py-func docutils literal notranslate"><span class="pre">metrics.auc_score</span></code> and
<a class="reference internal" href="../modules/generated/sklearn.metrics.average_precision_score.html#sklearn.metrics.average_precision_score" title="sklearn.metrics.average_precision_score"><code class="xref py py-func docutils literal notranslate"><span class="pre">metrics.average_precision_score</span></code></a> convenience functions by <a class="reference external" href="https://amueller.github.io/">Andreas
Müller</a>.</p></li>
<li><p>Improved sparse matrix support in the <a class="reference internal" href="../modules/feature_selection.html#feature-selection"><span class="std std-ref">Feature selection</span></a>
module by <a class="reference external" href="https://amueller.github.io/">Andreas Müller</a>.</p></li>
<li><p>New word boundaries-aware character n-gram analyzer for the
<a class="reference internal" href="../modules/feature_extraction.html#text-feature-extraction"><span class="std std-ref">Text feature extraction</span></a> module by <a class="reference external" href="https://github.com/kernc">@kernc</a>.</p></li>
<li><p>Fixed bug in spectral clustering that led to single point clusters
by <a class="reference external" href="https://amueller.github.io/">Andreas Müller</a>.</p></li>
<li><p>In <a class="reference internal" href="../modules/generated/sklearn.feature_extraction.text.CountVectorizer.html#sklearn.feature_extraction.text.CountVectorizer" title="sklearn.feature_extraction.text.CountVectorizer"><code class="xref py py-class docutils literal notranslate"><span class="pre">feature_extraction.text.CountVectorizer</span></code></a>, added an option to
ignore infrequent words, <code class="docutils literal notranslate"><span class="pre">min_df</span></code> by <a class="reference external" href="https://amueller.github.io/">Andreas Müller</a>.</p></li>
<li><p>Add support for multiple targets in some linear models (ElasticNet, Lasso
and OrthogonalMatchingPursuit) by <a class="reference external" href="https://vene.ro/">Vlad Niculae</a> and
<a class="reference external" href="http://alexandre.gramfort.net">Alexandre Gramfort</a>.</p></li>
<li><p>Fixes in <code class="xref py py-class docutils literal notranslate"><span class="pre">decomposition.ProbabilisticPCA</span></code> score function by Wei Li.</p></li>
<li><p>Fixed feature importance computation in
<a class="reference internal" href="../modules/ensemble.html#gradient-boosting"><span class="std std-ref">Gradient Tree Boosting</span></a>.</p></li>
</ul>
</div>
<div class="section" id="api-changes-summary">
<h2>API changes summary<a class="headerlink" href="#api-changes-summary" title="Permalink to this headline">¶</a></h2>
<ul class="simple">
<li><p>The old <code class="docutils literal notranslate"><span class="pre">scikits.learn</span></code> package has disappeared; all code should import
from <code class="docutils literal notranslate"><span class="pre">sklearn</span></code> instead, which was introduced in 0.9.</p></li>
<li><p>In <a class="reference internal" href="../modules/generated/sklearn.metrics.roc_curve.html#sklearn.metrics.roc_curve" title="sklearn.metrics.roc_curve"><code class="xref py py-func docutils literal notranslate"><span class="pre">metrics.roc_curve</span></code></a>, the <code class="docutils literal notranslate"><span class="pre">thresholds</span></code> array is now returned
with it’s order reversed, in order to keep it consistent with the order
of the returned <code class="docutils literal notranslate"><span class="pre">fpr</span></code> and <code class="docutils literal notranslate"><span class="pre">tpr</span></code>.</p></li>
<li><p>In <code class="xref py py-class docutils literal notranslate"><span class="pre">hmm</span></code> objects, like <code class="xref py py-class docutils literal notranslate"><span class="pre">hmm.GaussianHMM</span></code>,
<code class="xref py py-class docutils literal notranslate"><span class="pre">hmm.MultinomialHMM</span></code>, etc., all parameters must be passed to the
object when initialising it and not through <code class="docutils literal notranslate"><span class="pre">fit</span></code>. Now <code class="docutils literal notranslate"><span class="pre">fit</span></code> will
only accept the data as an input parameter.</p></li>
<li><p>For all SVM classes, a faulty behavior of <code class="docutils literal notranslate"><span class="pre">gamma</span></code> was fixed. Previously,
the default gamma value was only computed the first time <code class="docutils literal notranslate"><span class="pre">fit</span></code> was called
and then stored. It is now recalculated on every call to <code class="docutils literal notranslate"><span class="pre">fit</span></code>.</p></li>
<li><p>All <code class="docutils literal notranslate"><span class="pre">Base</span></code> classes are now abstract meta classes so that they can not be
instantiated.</p></li>
<li><p><a class="reference internal" href="../modules/generated/sklearn.cluster.ward_tree.html#sklearn.cluster.ward_tree" title="sklearn.cluster.ward_tree"><code class="xref py py-func docutils literal notranslate"><span class="pre">cluster.ward_tree</span></code></a> now also returns the parent array. This is
necessary for early-stopping in which case the tree is not
completely built.</p></li>
<li><p>In <a class="reference internal" href="../modules/generated/sklearn.feature_extraction.text.CountVectorizer.html#sklearn.feature_extraction.text.CountVectorizer" title="sklearn.feature_extraction.text.CountVectorizer"><code class="xref py py-class docutils literal notranslate"><span class="pre">feature_extraction.text.CountVectorizer</span></code></a> the parameters
<code class="docutils literal notranslate"><span class="pre">min_n</span></code> and <code class="docutils literal notranslate"><span class="pre">max_n</span></code> were joined to the parameter <code class="docutils literal notranslate"><span class="pre">n_gram_range</span></code> to
enable grid-searching both at once.</p></li>
<li><p>In <a class="reference internal" href="../modules/generated/sklearn.feature_extraction.text.CountVectorizer.html#sklearn.feature_extraction.text.CountVectorizer" title="sklearn.feature_extraction.text.CountVectorizer"><code class="xref py py-class docutils literal notranslate"><span class="pre">feature_extraction.text.CountVectorizer</span></code></a>, words that appear
only in one document are now ignored by default. To reproduce
the previous behavior, set <code class="docutils literal notranslate"><span class="pre">min_df=1</span></code>.</p></li>
<li><p>Fixed API inconsistency: <a class="reference internal" href="../modules/generated/sklearn.linear_model.SGDClassifier.html#sklearn.linear_model.SGDClassifier.predict_proba" title="sklearn.linear_model.SGDClassifier.predict_proba"><code class="xref py py-meth docutils literal notranslate"><span class="pre">linear_model.SGDClassifier.predict_proba</span></code></a> now
returns 2d array when fit on two classes.</p></li>
<li><p>Fixed API inconsistency: <a class="reference internal" href="../modules/generated/sklearn.discriminant_analysis.QuadraticDiscriminantAnalysis.html#sklearn.discriminant_analysis.QuadraticDiscriminantAnalysis.decision_function" title="sklearn.discriminant_analysis.QuadraticDiscriminantAnalysis.decision_function"><code class="xref py py-meth docutils literal notranslate"><span class="pre">discriminant_analysis.QuadraticDiscriminantAnalysis.decision_function</span></code></a>
and <a class="reference internal" href="../modules/generated/sklearn.discriminant_analysis.LinearDiscriminantAnalysis.html#sklearn.discriminant_analysis.LinearDiscriminantAnalysis.decision_function" title="sklearn.discriminant_analysis.LinearDiscriminantAnalysis.decision_function"><code class="xref py py-meth docutils literal notranslate"><span class="pre">discriminant_analysis.LinearDiscriminantAnalysis.decision_function</span></code></a> now return 1d arrays
when fit on two classes.</p></li>
<li><p>Grid of alphas used for fitting <a class="reference internal" href="../modules/generated/sklearn.linear_model.LassoCV.html#sklearn.linear_model.LassoCV" title="sklearn.linear_model.LassoCV"><code class="xref py py-class docutils literal notranslate"><span class="pre">linear_model.LassoCV</span></code></a> and
<a class="reference internal" href="../modules/generated/sklearn.linear_model.ElasticNetCV.html#sklearn.linear_model.ElasticNetCV" title="sklearn.linear_model.ElasticNetCV"><code class="xref py py-class docutils literal notranslate"><span class="pre">linear_model.ElasticNetCV</span></code></a> is now stored
in the attribute <code class="docutils literal notranslate"><span class="pre">alphas_</span></code> rather than overriding the init parameter
<code class="docutils literal notranslate"><span class="pre">alphas</span></code>.</p></li>
<li><p>Linear models when alpha is estimated by cross-validation store
the estimated value in the <code class="docutils literal notranslate"><span class="pre">alpha_</span></code> attribute rather than just
<code class="docutils literal notranslate"><span class="pre">alpha</span></code> or <code class="docutils literal notranslate"><span class="pre">best_alpha</span></code>.</p></li>
<li><p><a class="reference internal" href="../modules/generated/sklearn.ensemble.GradientBoostingClassifier.html#sklearn.ensemble.GradientBoostingClassifier" title="sklearn.ensemble.GradientBoostingClassifier"><code class="xref py py-class docutils literal notranslate"><span class="pre">ensemble.GradientBoostingClassifier</span></code></a> now supports
<a class="reference internal" href="../modules/generated/sklearn.ensemble.GradientBoostingClassifier.html#sklearn.ensemble.GradientBoostingClassifier.staged_predict_proba" title="sklearn.ensemble.GradientBoostingClassifier.staged_predict_proba"><code class="xref py py-meth docutils literal notranslate"><span class="pre">ensemble.GradientBoostingClassifier.staged_predict_proba</span></code></a>, and
<a class="reference internal" href="../modules/generated/sklearn.ensemble.GradientBoostingClassifier.html#sklearn.ensemble.GradientBoostingClassifier.staged_predict" title="sklearn.ensemble.GradientBoostingClassifier.staged_predict"><code class="xref py py-meth docutils literal notranslate"><span class="pre">ensemble.GradientBoostingClassifier.staged_predict</span></code></a>.</p></li>
<li><p><code class="xref py py-class docutils literal notranslate"><span class="pre">svm.sparse.SVC</span></code> and other sparse SVM classes are now deprecated.
The all classes in the <a class="reference internal" href="../modules/svm.html#svm"><span class="std std-ref">Support Vector Machines</span></a> module now automatically select the
sparse or dense representation base on the input.</p></li>
<li><p>All clustering algorithms now interpret the array <code class="docutils literal notranslate"><span class="pre">X</span></code> given to <code class="docutils literal notranslate"><span class="pre">fit</span></code> as
input data, in particular <a class="reference internal" href="../modules/generated/sklearn.cluster.SpectralClustering.html#sklearn.cluster.SpectralClustering" title="sklearn.cluster.SpectralClustering"><code class="xref py py-class docutils literal notranslate"><span class="pre">cluster.SpectralClustering</span></code></a> and
<a class="reference internal" href="../modules/generated/sklearn.cluster.AffinityPropagation.html#sklearn.cluster.AffinityPropagation" title="sklearn.cluster.AffinityPropagation"><code class="xref py py-class docutils literal notranslate"><span class="pre">cluster.AffinityPropagation</span></code></a> which previously expected affinity matrices.</p></li>
<li><p>For clustering algorithms that take the desired number of clusters as a parameter,
this parameter is now called <code class="docutils literal notranslate"><span class="pre">n_clusters</span></code>.</p></li>
</ul>
</div>
<div class="section" id="id2">
<h2>People<a class="headerlink" href="#id2" title="Permalink to this headline">¶</a></h2>
<blockquote>
<div><ul class="simple">
<li><p>267 <a class="reference external" href="https://amueller.github.io/">Andreas Müller</a></p></li>
<li><p>94 <a class="reference external" href="http://www.montefiore.ulg.ac.be/~glouppe/">Gilles Louppe</a></p></li>
<li><p>89 <a class="reference external" href="http://gael-varoquaux.info">Gael Varoquaux</a></p></li>
<li><p>79 <a class="reference external" href="https://sites.google.com/site/peterprettenhofer/">Peter Prettenhofer</a></p></li>
<li><p>60 <a class="reference external" href="http://www.mblondel.org">Mathieu Blondel</a></p></li>
<li><p>57 <a class="reference external" href="http://alexandre.gramfort.net">Alexandre Gramfort</a></p></li>
<li><p>52 <a class="reference external" href="https://vene.ro/">Vlad Niculae</a></p></li>
<li><p>45 <a class="reference external" href="https://github.com/larsmans">Lars Buitinck</a></p></li>
<li><p>44 Nelle Varoquaux</p></li>
<li><p>37 <a class="reference external" href="https://github.com/jaquesgrobler">Jaques Grobler</a></p></li>
<li><p>30 Alexis Mignon</p></li>
<li><p>30 Immanuel Bayer</p></li>
<li><p>27 <a class="reference external" href="https://twitter.com/ogrisel">Olivier Grisel</a></p></li>
<li><p>16 Subhodeep Moitra</p></li>
<li><p>13 Yannick Schwartz</p></li>
<li><p>12 <a class="reference external" href="https://github.com/kernc">@kernc</a></p></li>
<li><p>11 <a class="reference external" href="https://github.com/VirgileFritsch">Virgile Fritsch</a></p></li>
<li><p>9 Daniel Duckworth</p></li>
<li><p>9 <a class="reference external" href="http://fa.bianp.net">Fabian Pedregosa</a></p></li>
<li><p>9 <a class="reference external" href="https://twitter.com/robertlayton">Robert Layton</a></p></li>
<li><p>8 John Benediktsson</p></li>
<li><p>7 Marko Burjek</p></li>
<li><p>5 <a class="reference external" href="https://twitter.com/npinto">Nicolas Pinto</a></p></li>
<li><p>4 Alexandre Abraham</p></li>
<li><p>4 <a class="reference external" href="https://staff.washington.edu/jakevdp/">Jake Vanderplas</a></p></li>
<li><p>3 <a class="reference external" href="http://personal.ee.surrey.ac.uk/Personal/B.Holt">Brian Holt</a></p></li>
<li><p>3 <a class="reference external" href="https://sites.google.com/site/duchesnay/home">Edouard Duchesnay</a></p></li>
<li><p>3 Florian Hoenig</p></li>
<li><p>3 flyingimmidev</p></li>
<li><p>2 Francois Savard</p></li>
<li><p>2 Hannes Schulz</p></li>
<li><p>2 Peter Welinder</p></li>
<li><p>2 <a class="reference external" href="http://www.onerussian.com/">Yaroslav Halchenko</a></p></li>
<li><p>2 Wei Li</p></li>
<li><p>1 Alex Companioni</p></li>
<li><p>1 Brandyn A. White</p></li>
<li><p>1 Bussonnier Matthias</p></li>
<li><p>1 Charles-Pierre Astolfi</p></li>
<li><p>1 Dan O’Huiginn</p></li>
<li><p>1 David Cournapeau</p></li>
<li><p>1 Keith Goodman</p></li>
<li><p>1 Ludwig Schwardt</p></li>
<li><p>1 Olivier Hervieu</p></li>
<li><p>1 Sergio Medina</p></li>
<li><p>1 Shiqiao Du</p></li>
<li><p>1 Tim Sheerman-Chase</p></li>
<li><p>1 buguen</p></li>
</ul>
</div></blockquote>
</div>
</div>
<div class="section" id="version-0-11">
<span id="changes-0-11"></span><h1>Version 0.11<a class="headerlink" href="#version-0-11" title="Permalink to this headline">¶</a></h1>
<p><strong>May 7, 2012</strong></p>
<div class="section" id="id3">
<h2>Changelog<a class="headerlink" href="#id3" title="Permalink to this headline">¶</a></h2>
<div class="section" id="highlights">
<h3>Highlights<a class="headerlink" href="#highlights" title="Permalink to this headline">¶</a></h3>
<ul class="simple">
<li><p>Gradient boosted regression trees (<a class="reference internal" href="../modules/ensemble.html#gradient-boosting"><span class="std std-ref">Gradient Tree Boosting</span></a>)
for classification and regression by <a class="reference external" href="https://sites.google.com/site/peterprettenhofer/">Peter Prettenhofer</a>
and <a class="reference external" href="https://twitter.com/scottblanc">Scott White</a> .</p></li>
<li><p>Simple dict-based feature loader with support for categorical variables
(<a class="reference internal" href="../modules/generated/sklearn.feature_extraction.DictVectorizer.html#sklearn.feature_extraction.DictVectorizer" title="sklearn.feature_extraction.DictVectorizer"><code class="xref py py-class docutils literal notranslate"><span class="pre">feature_extraction.DictVectorizer</span></code></a>) by <a class="reference external" href="https://github.com/larsmans">Lars Buitinck</a>.</p></li>
<li><p>Added Matthews correlation coefficient (<a class="reference internal" href="../modules/generated/sklearn.metrics.matthews_corrcoef.html#sklearn.metrics.matthews_corrcoef" title="sklearn.metrics.matthews_corrcoef"><code class="xref py py-func docutils literal notranslate"><span class="pre">metrics.matthews_corrcoef</span></code></a>)
and added macro and micro average options to
<a class="reference internal" href="../modules/generated/sklearn.metrics.precision_score.html#sklearn.metrics.precision_score" title="sklearn.metrics.precision_score"><code class="xref py py-func docutils literal notranslate"><span class="pre">metrics.precision_score</span></code></a>, <a class="reference internal" href="../modules/generated/sklearn.metrics.recall_score.html#sklearn.metrics.recall_score" title="sklearn.metrics.recall_score"><code class="xref py py-func docutils literal notranslate"><span class="pre">metrics.recall_score</span></code></a> and
<a class="reference internal" href="../modules/generated/sklearn.metrics.f1_score.html#sklearn.metrics.f1_score" title="sklearn.metrics.f1_score"><code class="xref py py-func docutils literal notranslate"><span class="pre">metrics.f1_score</span></code></a> by <a class="reference external" href="https://www.mit.edu/~satra/">Satrajit Ghosh</a>.</p></li>
<li><p><a class="reference internal" href="../modules/grid_search.html#out-of-bag"><span class="std std-ref">Out of Bag Estimates</span></a> of generalization error for <a class="reference internal" href="../modules/ensemble.html#ensemble"><span class="std std-ref">Ensemble methods</span></a>
by <a class="reference external" href="https://amueller.github.io/">Andreas Müller</a>.</p></li>
<li><p>Randomized sparse linear models for feature
selection, by <a class="reference external" href="http://alexandre.gramfort.net">Alexandre Gramfort</a> and <a class="reference external" href="http://gael-varoquaux.info">Gael Varoquaux</a></p></li>
<li><p><a class="reference internal" href="../modules/label_propagation.html#label-propagation"><span class="std std-ref">Label Propagation</span></a> for semi-supervised learning, by Clay
Woolam. <strong>Note</strong> the semi-supervised API is still work in progress,
and may change.</p></li>
<li><p>Added BIC/AIC model selection to classical <a class="reference internal" href="../modules/mixture.html#gmm"><span class="std std-ref">Gaussian mixture models</span></a> and unified
the API with the remainder of scikit-learn, by <a class="reference external" href="https://team.inria.fr/parietal/bertrand-thirions-page">Bertrand Thirion</a></p></li>
<li><p>Added <code class="xref py py-class docutils literal notranslate"><span class="pre">sklearn.cross_validation.StratifiedShuffleSplit</span></code>, which is
a <code class="xref py py-class docutils literal notranslate"><span class="pre">sklearn.cross_validation.ShuffleSplit</span></code> with balanced splits,
by Yannick Schwartz.</p></li>
<li><p><a class="reference internal" href="../modules/generated/sklearn.neighbors.NearestCentroid.html#sklearn.neighbors.NearestCentroid" title="sklearn.neighbors.NearestCentroid"><code class="xref py py-class docutils literal notranslate"><span class="pre">sklearn.neighbors.NearestCentroid</span></code></a> classifier added, along with a
<code class="docutils literal notranslate"><span class="pre">shrink_threshold</span></code> parameter, which implements <strong>shrunken centroid
classification</strong>, by <a class="reference external" href="https://twitter.com/robertlayton">Robert Layton</a>.</p></li>
</ul>
</div>
<div class="section" id="other-changes">
<h3>Other changes<a class="headerlink" href="#other-changes" title="Permalink to this headline">¶</a></h3>
<ul class="simple">
<li><p>Merged dense and sparse implementations of <a class="reference internal" href="../modules/sgd.html#sgd"><span class="std std-ref">Stochastic Gradient Descent</span></a> module and
exposed utility extension types for sequential
datasets <code class="docutils literal notranslate"><span class="pre">seq_dataset</span></code> and weight vectors <code class="docutils literal notranslate"><span class="pre">weight_vector</span></code>
by <a class="reference external" href="https://sites.google.com/site/peterprettenhofer/">Peter Prettenhofer</a>.</p></li>
<li><p>Added <code class="docutils literal notranslate"><span class="pre">partial_fit</span></code> (support for online/minibatch learning) and
warm_start to the <a class="reference internal" href="../modules/sgd.html#sgd"><span class="std std-ref">Stochastic Gradient Descent</span></a> module by <a class="reference external" href="http://www.mblondel.org">Mathieu Blondel</a>.</p></li>
<li><p>Dense and sparse implementations of <a class="reference internal" href="../modules/svm.html#svm"><span class="std std-ref">Support Vector Machines</span></a> classes and
<a class="reference internal" href="../modules/generated/sklearn.linear_model.LogisticRegression.html#sklearn.linear_model.LogisticRegression" title="sklearn.linear_model.LogisticRegression"><code class="xref py py-class docutils literal notranslate"><span class="pre">linear_model.LogisticRegression</span></code></a> merged by <a class="reference external" href="https://github.com/larsmans">Lars Buitinck</a>.</p></li>
<li><p>Regressors can now be used as base estimator in the <a class="reference internal" href="../modules/multiclass.html#multiclass"><span class="std std-ref">Multiclass and multilabel algorithms</span></a>
module by <a class="reference external" href="http://www.mblondel.org">Mathieu Blondel</a>.</p></li>
<li><p>Added n_jobs option to <code class="xref py py-func docutils literal notranslate"><span class="pre">metrics.pairwise.pairwise_distances</span></code>
and <a class="reference internal" href="../modules/generated/sklearn.metrics.pairwise.pairwise_kernels.html#sklearn.metrics.pairwise.pairwise_kernels" title="sklearn.metrics.pairwise.pairwise_kernels"><code class="xref py py-func docutils literal notranslate"><span class="pre">metrics.pairwise.pairwise_kernels</span></code></a> for parallel computation,
by <a class="reference external" href="http://www.mblondel.org">Mathieu Blondel</a>.</p></li>
<li><p><a class="reference internal" href="../modules/clustering.html#k-means"><span class="std std-ref">K-means</span></a> can now be run in parallel, using the <code class="docutils literal notranslate"><span class="pre">n_jobs</span></code> argument
to either <a class="reference internal" href="../modules/clustering.html#k-means"><span class="std std-ref">K-means</span></a> or <code class="xref py py-class docutils literal notranslate"><span class="pre">KMeans</span></code>, by <a class="reference external" href="https://twitter.com/robertlayton">Robert Layton</a>.</p></li>
<li><p>Improved <a class="reference internal" href="../modules/cross_validation.html#cross-validation"><span class="std std-ref">Cross-validation: evaluating estimator performance</span></a> and <a class="reference internal" href="../modules/grid_search.html#grid-search"><span class="std std-ref">Tuning the hyper-parameters of an estimator</span></a> documentation
and introduced the new <code class="xref py py-func docutils literal notranslate"><span class="pre">cross_validation.train_test_split</span></code>
helper function by <a class="reference external" href="https://twitter.com/ogrisel">Olivier Grisel</a></p></li>
<li><p><a class="reference internal" href="../modules/generated/sklearn.svm.SVC.html#sklearn.svm.SVC" title="sklearn.svm.SVC"><code class="xref py py-class docutils literal notranslate"><span class="pre">svm.SVC</span></code></a> members <code class="docutils literal notranslate"><span class="pre">coef_</span></code> and <code class="docutils literal notranslate"><span class="pre">intercept_</span></code> changed sign for
consistency with <code class="docutils literal notranslate"><span class="pre">decision_function</span></code>; for <code class="docutils literal notranslate"><span class="pre">kernel==linear</span></code>,
<code class="docutils literal notranslate"><span class="pre">coef_</span></code> was fixed in the one-vs-one case, by <a class="reference external" href="https://amueller.github.io/">Andreas Müller</a>.</p></li>
<li><p>Performance improvements to efficient leave-one-out cross-validated
Ridge regression, esp. for the <code class="docutils literal notranslate"><span class="pre">n_samples</span> <span class="pre">></span> <span class="pre">n_features</span></code> case, in
<a class="reference internal" href="../modules/generated/sklearn.linear_model.RidgeCV.html#sklearn.linear_model.RidgeCV" title="sklearn.linear_model.RidgeCV"><code class="xref py py-class docutils literal notranslate"><span class="pre">linear_model.RidgeCV</span></code></a>, by Reuben Fletcher-Costin.</p></li>
<li><p>Refactoring and simplification of the <a class="reference internal" href="../modules/feature_extraction.html#text-feature-extraction"><span class="std std-ref">Text feature extraction</span></a>
API and fixed a bug that caused possible negative IDF,
by <a class="reference external" href="https://twitter.com/ogrisel">Olivier Grisel</a>.</p></li>
<li><p>Beam pruning option in <code class="xref py py-class docutils literal notranslate"><span class="pre">_BaseHMM</span></code> module has been removed since it
is difficult to Cythonize. If you are interested in contributing a Cython
version, you can use the python version in the git history as a reference.</p></li>
<li><p>Classes in <a class="reference internal" href="../modules/neighbors.html#neighbors"><span class="std std-ref">Nearest Neighbors</span></a> now support arbitrary Minkowski metric for
nearest neighbors searches. The metric can be specified by argument <code class="docutils literal notranslate"><span class="pre">p</span></code>.</p></li>
</ul>
</div>
</div>
<div class="section" id="id4">
<h2>API changes summary<a class="headerlink" href="#id4" title="Permalink to this headline">¶</a></h2>
<ul class="simple">
<li><p><code class="xref py py-class docutils literal notranslate"><span class="pre">covariance.EllipticEnvelop</span></code> is now deprecated - Please use <a class="reference internal" href="../modules/generated/sklearn.covariance.EllipticEnvelope.html#sklearn.covariance.EllipticEnvelope" title="sklearn.covariance.EllipticEnvelope"><code class="xref py py-class docutils literal notranslate"><span class="pre">covariance.EllipticEnvelope</span></code></a>
instead.</p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">NeighborsClassifier</span></code> and <code class="docutils literal notranslate"><span class="pre">NeighborsRegressor</span></code> are gone in the module
<a class="reference internal" href="../modules/neighbors.html#neighbors"><span class="std std-ref">Nearest Neighbors</span></a>. Use the classes <code class="xref py py-class docutils literal notranslate"><span class="pre">KNeighborsClassifier</span></code>,
<code class="xref py py-class docutils literal notranslate"><span class="pre">RadiusNeighborsClassifier</span></code>, <code class="xref py py-class docutils literal notranslate"><span class="pre">KNeighborsRegressor</span></code>
and/or <code class="xref py py-class docutils literal notranslate"><span class="pre">RadiusNeighborsRegressor</span></code> instead.</p></li>
<li><p>Sparse classes in the <a class="reference internal" href="../modules/sgd.html#sgd"><span class="std std-ref">Stochastic Gradient Descent</span></a> module are now deprecated.</p></li>
<li><p>In <code class="xref py py-class docutils literal notranslate"><span class="pre">mixture.GMM</span></code>, <code class="xref py py-class docutils literal notranslate"><span class="pre">mixture.DPGMM</span></code> and <code class="xref py py-class docutils literal notranslate"><span class="pre">mixture.VBGMM</span></code>,
parameters must be passed to an object when initialising it and not through
<code class="docutils literal notranslate"><span class="pre">fit</span></code>. Now <code class="docutils literal notranslate"><span class="pre">fit</span></code> will only accept the data as an input parameter.</p></li>
<li><p>methods <code class="docutils literal notranslate"><span class="pre">rvs</span></code> and <code class="docutils literal notranslate"><span class="pre">decode</span></code> in <code class="xref py py-class docutils literal notranslate"><span class="pre">GMM</span></code> module are now deprecated.
<code class="docutils literal notranslate"><span class="pre">sample</span></code> and <code class="docutils literal notranslate"><span class="pre">score</span></code> or <code class="docutils literal notranslate"><span class="pre">predict</span></code> should be used instead.</p></li>
<li><p>attribute <code class="docutils literal notranslate"><span class="pre">_scores</span></code> and <code class="docutils literal notranslate"><span class="pre">_pvalues</span></code> in univariate feature selection
objects are now deprecated.
<code class="docutils literal notranslate"><span class="pre">scores_</span></code> or <code class="docutils literal notranslate"><span class="pre">pvalues_</span></code> should be used instead.</p></li>
<li><p>In <code class="xref py py-class docutils literal notranslate"><span class="pre">LogisticRegression</span></code>, <code class="xref py py-class docutils literal notranslate"><span class="pre">LinearSVC</span></code>, <code class="xref py py-class docutils literal notranslate"><span class="pre">SVC</span></code> and
<code class="xref py py-class docutils literal notranslate"><span class="pre">NuSVC</span></code>, the <code class="docutils literal notranslate"><span class="pre">class_weight</span></code> parameter is now an initialization
parameter, not a parameter to fit. This makes grid searches
over this parameter possible.</p></li>
<li><p>LFW <code class="docutils literal notranslate"><span class="pre">data</span></code> is now always shape <code class="docutils literal notranslate"><span class="pre">(n_samples,</span> <span class="pre">n_features)</span></code> to be
consistent with the Olivetti faces dataset. Use <code class="docutils literal notranslate"><span class="pre">images</span></code> and
<code class="docutils literal notranslate"><span class="pre">pairs</span></code> attribute to access the natural images shapes instead.</p></li>
<li><p>In <a class="reference internal" href="../modules/generated/sklearn.svm.LinearSVC.html#sklearn.svm.LinearSVC" title="sklearn.svm.LinearSVC"><code class="xref py py-class docutils literal notranslate"><span class="pre">svm.LinearSVC</span></code></a>, the meaning of the <code class="docutils literal notranslate"><span class="pre">multi_class</span></code> parameter
changed. Options now are <code class="docutils literal notranslate"><span class="pre">'ovr'</span></code> and <code class="docutils literal notranslate"><span class="pre">'crammer_singer'</span></code>, with
<code class="docutils literal notranslate"><span class="pre">'ovr'</span></code> being the default. This does not change the default behavior
but hopefully is less confusing.</p></li>
<li><p>Class <code class="xref py py-class docutils literal notranslate"><span class="pre">feature_selection.text.Vectorizer</span></code> is deprecated and
replaced by <code class="xref py py-class docutils literal notranslate"><span class="pre">feature_selection.text.TfidfVectorizer</span></code>.</p></li>
<li><p>The preprocessor / analyzer nested structure for text feature
extraction has been removed. All those features are
now directly passed as flat constructor arguments
to <code class="xref py py-class docutils literal notranslate"><span class="pre">feature_selection.text.TfidfVectorizer</span></code> and
<code class="xref py py-class docutils literal notranslate"><span class="pre">feature_selection.text.CountVectorizer</span></code>, in particular the
following parameters are now used:</p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">analyzer</span></code> can be <code class="docutils literal notranslate"><span class="pre">'word'</span></code> or <code class="docutils literal notranslate"><span class="pre">'char'</span></code> to switch the default
analysis scheme, or use a specific python callable (as previously).</p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">tokenizer</span></code> and <code class="docutils literal notranslate"><span class="pre">preprocessor</span></code> have been introduced to make it
still possible to customize those steps with the new API.</p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">input</span></code> explicitly control how to interpret the sequence passed to
<code class="docutils literal notranslate"><span class="pre">fit</span></code> and <code class="docutils literal notranslate"><span class="pre">predict</span></code>: filenames, file objects or direct (byte or
Unicode) strings.</p></li>
<li><p>charset decoding is explicit and strict by default.</p></li>
<li><p>the <code class="docutils literal notranslate"><span class="pre">vocabulary</span></code>, fitted or not is now stored in the
<code class="docutils literal notranslate"><span class="pre">vocabulary_</span></code> attribute to be consistent with the project
conventions.</p></li>
<li><p>Class <code class="xref py py-class docutils literal notranslate"><span class="pre">feature_selection.text.TfidfVectorizer</span></code> now derives directly
from <code class="xref py py-class docutils literal notranslate"><span class="pre">feature_selection.text.CountVectorizer</span></code> to make grid
search trivial.</p></li>
<li><p>methods <code class="docutils literal notranslate"><span class="pre">rvs</span></code> in <code class="xref py py-class docutils literal notranslate"><span class="pre">_BaseHMM</span></code> module are now deprecated.
<code class="docutils literal notranslate"><span class="pre">sample</span></code> should be used instead.</p></li>
<li><p>Beam pruning option in <code class="xref py py-class docutils literal notranslate"><span class="pre">_BaseHMM</span></code> module is removed since it is
difficult to be Cythonized. If you are interested, you can look in the
history codes by git.</p></li>
<li><p>The SVMlight format loader now supports files with both zero-based and
one-based column indices, since both occur “in the wild”.</p></li>
<li><p>Arguments in class <code class="xref py py-class docutils literal notranslate"><span class="pre">ShuffleSplit</span></code> are now consistent with
<code class="xref py py-class docutils literal notranslate"><span class="pre">StratifiedShuffleSplit</span></code>. Arguments <code class="docutils literal notranslate"><span class="pre">test_fraction</span></code> and
<code class="docutils literal notranslate"><span class="pre">train_fraction</span></code> are deprecated and renamed to <code class="docutils literal notranslate"><span class="pre">test_size</span></code> and
<code class="docutils literal notranslate"><span class="pre">train_size</span></code> and can accept both <code class="docutils literal notranslate"><span class="pre">float</span></code> and <code class="docutils literal notranslate"><span class="pre">int</span></code>.</p></li>
<li><p>Arguments in class <code class="xref py py-class docutils literal notranslate"><span class="pre">Bootstrap</span></code> are now consistent with
<code class="xref py py-class docutils literal notranslate"><span class="pre">StratifiedShuffleSplit</span></code>. Arguments <code class="docutils literal notranslate"><span class="pre">n_test</span></code> and
<code class="docutils literal notranslate"><span class="pre">n_train</span></code> are deprecated and renamed to <code class="docutils literal notranslate"><span class="pre">test_size</span></code> and
<code class="docutils literal notranslate"><span class="pre">train_size</span></code> and can accept both <code class="docutils literal notranslate"><span class="pre">float</span></code> and <code class="docutils literal notranslate"><span class="pre">int</span></code>.</p></li>
<li><p>Argument <code class="docutils literal notranslate"><span class="pre">p</span></code> added to classes in <a class="reference internal" href="../modules/neighbors.html#neighbors"><span class="std std-ref">Nearest Neighbors</span></a> to specify an
arbitrary Minkowski metric for nearest neighbors searches.</p></li>
</ul>
</div>
<div class="section" id="id5">
<h2>People<a class="headerlink" href="#id5" title="Permalink to this headline">¶</a></h2>
<blockquote>
<div><ul class="simple">
<li><p>282 <a class="reference external" href="https://amueller.github.io/">Andreas Müller</a></p></li>
<li><p>239 <a class="reference external" href="https://sites.google.com/site/peterprettenhofer/">Peter Prettenhofer</a></p></li>
<li><p>198 <a class="reference external" href="http://gael-varoquaux.info">Gael Varoquaux</a></p></li>
<li><p>129 <a class="reference external" href="https://twitter.com/ogrisel">Olivier Grisel</a></p></li>
<li><p>114 <a class="reference external" href="http://www.mblondel.org">Mathieu Blondel</a></p></li>
<li><p>103 Clay Woolam</p></li>
<li><p>96 <a class="reference external" href="https://github.com/larsmans">Lars Buitinck</a></p></li>
<li><p>88 <a class="reference external" href="https://github.com/jaquesgrobler">Jaques Grobler</a></p></li>
<li><p>82 <a class="reference external" href="http://alexandre.gramfort.net">Alexandre Gramfort</a></p></li>
<li><p>50 <a class="reference external" href="https://team.inria.fr/parietal/bertrand-thirions-page">Bertrand Thirion</a></p></li>
<li><p>42 <a class="reference external" href="https://twitter.com/robertlayton">Robert Layton</a></p></li>
<li><p>28 flyingimmidev</p></li>
<li><p>26 <a class="reference external" href="https://staff.washington.edu/jakevdp/">Jake Vanderplas</a></p></li>
<li><p>26 Shiqiao Du</p></li>
<li><p>21 <a class="reference external" href="https://www.mit.edu/~satra/">Satrajit Ghosh</a></p></li>
<li><p>17 <a class="reference external" href="https://davidmarek.cz/">David Marek</a></p></li>
<li><p>17 <a class="reference external" href="http://www.montefiore.ulg.ac.be/~glouppe/">Gilles Louppe</a></p></li>
<li><p>14 <a class="reference external" href="https://vene.ro/">Vlad Niculae</a></p></li>
<li><p>11 Yannick Schwartz</p></li>
<li><p>10 <a class="reference external" href="http://fa.bianp.net">Fabian Pedregosa</a></p></li>
<li><p>9 fcostin</p></li>
<li><p>7 Nick Wilson</p></li>
<li><p>5 Adrien Gaidon</p></li>
<li><p>5 <a class="reference external" href="https://twitter.com/npinto">Nicolas Pinto</a></p></li>
<li><p>4 <a class="reference external" href="http://www-etud.iro.umontreal.ca/~wardefar/">David Warde-Farley</a></p></li>
<li><p>5 Nelle Varoquaux</p></li>
<li><p>5 Emmanuelle Gouillart</p></li>
<li><p>3 Joonas Sillanpää</p></li>
<li><p>3 Paolo Losi</p></li>
<li><p>2 Charles McCarthy</p></li>
<li><p>2 Roy Hyunjin Han</p></li>
<li><p>2 Scott White</p></li>
<li><p>2 ibayer</p></li>
<li><p>1 Brandyn White</p></li>
<li><p>1 Carlos Scheidegger</p></li>
<li><p>1 Claire Revillet</p></li>
<li><p>1 Conrad Lee</p></li>
<li><p>1 <a class="reference external" href="https://sites.google.com/site/duchesnay/home">Edouard Duchesnay</a></p></li>
<li><p>1 Jan Hendrik Metzen</p></li>
<li><p>1 Meng Xinfan</p></li>
<li><p>1 <a class="reference external" href="https://www.zinkov.com/">Rob Zinkov</a></p></li>
<li><p>1 Shiqiao</p></li>
<li><p>1 Udi Weinsberg</p></li>
<li><p>1 Virgile Fritsch</p></li>
<li><p>1 Xinfan Meng</p></li>
<li><p>1 Yaroslav Halchenko</p></li>
<li><p>1 jansoe</p></li>
<li><p>1 Leon Palafox</p></li>
</ul>
</div></blockquote>
</div>
</div>
<div class="section" id="version-0-10">
<span id="changes-0-10"></span><h1>Version 0.10<a class="headerlink" href="#version-0-10" title="Permalink to this headline">¶</a></h1>
<p><strong>January 11, 2012</strong></p>
<div class="section" id="id6">
<h2>Changelog<a class="headerlink" href="#id6" title="Permalink to this headline">¶</a></h2>
<ul class="simple">
<li><p>Python 2.5 compatibility was dropped; the minimum Python version needed
to use scikit-learn is now 2.6.</p></li>
<li><p><a class="reference internal" href="../modules/covariance.html#sparse-inverse-covariance"><span class="std std-ref">Sparse inverse covariance</span></a> estimation using the graph Lasso, with
associated cross-validated estimator, by <a class="reference external" href="http://gael-varoquaux.info">Gael Varoquaux</a></p></li>
<li><p>New <a class="reference internal" href="../modules/tree.html#tree"><span class="std std-ref">Tree</span></a> module by <a class="reference external" href="http://personal.ee.surrey.ac.uk/Personal/B.Holt">Brian Holt</a>, <a class="reference external" href="https://sites.google.com/site/peterprettenhofer/">Peter Prettenhofer</a>,
<a class="reference external" href="https://www.mit.edu/~satra/">Satrajit Ghosh</a> and <a class="reference external" href="http://www.montefiore.ulg.ac.be/~glouppe/">Gilles Louppe</a>. The module comes with complete
documentation and examples.</p></li>
<li><p>Fixed a bug in the RFE module by <a class="reference external" href="http://www.montefiore.ulg.ac.be/~glouppe/">Gilles Louppe</a> (issue #378).</p></li>
<li><p>Fixed a memory leak in <a class="reference internal" href="../modules/svm.html#svm"><span class="std std-ref">Support Vector Machines</span></a> module by <a class="reference external" href="http://personal.ee.surrey.ac.uk/Personal/B.Holt">Brian Holt</a> (issue #367).</p></li>
<li><p>Faster tests by <a class="reference external" href="http://fa.bianp.net">Fabian Pedregosa</a> and others.</p></li>
<li><p>Silhouette Coefficient cluster analysis evaluation metric added as
<a class="reference internal" href="../modules/generated/sklearn.metrics.silhouette_score.html#sklearn.metrics.silhouette_score" title="sklearn.metrics.silhouette_score"><code class="xref py py-func docutils literal notranslate"><span class="pre">sklearn.metrics.silhouette_score</span></code></a> by Robert Layton.</p></li>
<li><p>Fixed a bug in <a class="reference internal" href="../modules/clustering.html#k-means"><span class="std std-ref">K-means</span></a> in the handling of the <code class="docutils literal notranslate"><span class="pre">n_init</span></code> parameter:
the clustering algorithm used to be run <code class="docutils literal notranslate"><span class="pre">n_init</span></code> times but the last
solution was retained instead of the best solution by <a class="reference external" href="https://twitter.com/ogrisel">Olivier Grisel</a>.</p></li>
<li><p>Minor refactoring in <a class="reference internal" href="../modules/sgd.html#sgd"><span class="std std-ref">Stochastic Gradient Descent</span></a> module; consolidated dense and sparse
predict methods; Enhanced test time performance by converting model
parameters to fortran-style arrays after fitting (only multi-class).</p></li>
<li><p>Adjusted Mutual Information metric added as
<a class="reference internal" href="../modules/generated/sklearn.metrics.adjusted_mutual_info_score.html#sklearn.metrics.adjusted_mutual_info_score" title="sklearn.metrics.adjusted_mutual_info_score"><code class="xref py py-func docutils literal notranslate"><span class="pre">sklearn.metrics.adjusted_mutual_info_score</span></code></a> by Robert Layton.</p></li>
<li><p>Models like SVC/SVR/LinearSVC/LogisticRegression from libsvm/liblinear
now support scaling of C regularization parameter by the number of
samples by <a class="reference external" href="http://alexandre.gramfort.net">Alexandre Gramfort</a>.</p></li>
<li><p>New <a class="reference internal" href="../modules/ensemble.html#ensemble"><span class="std std-ref">Ensemble Methods</span></a> module by <a class="reference external" href="http://www.montefiore.ulg.ac.be/~glouppe/">Gilles Louppe</a> and
<a class="reference external" href="http://personal.ee.surrey.ac.uk/Personal/B.Holt">Brian Holt</a>. The module comes with the random forest algorithm and the
extra-trees method, along with documentation and examples.</p></li>
<li><p><a class="reference internal" href="../modules/outlier_detection.html#outlier-detection"><span class="std std-ref">Novelty and Outlier Detection</span></a>: outlier and novelty detection, by
<a class="reference external" href="https://github.com/VirgileFritsch">Virgile Fritsch</a>.</p></li>
<li><p><a class="reference internal" href="../modules/kernel_approximation.html#kernel-approximation"><span class="std std-ref">Kernel Approximation</span></a>: a transform implementing kernel
approximation for fast SGD on non-linear kernels by
<a class="reference external" href="https://amueller.github.io/">Andreas Müller</a>.</p></li>
<li><p>Fixed a bug due to atom swapping in <a class="reference internal" href="../modules/linear_model.html#omp"><span class="std std-ref">Orthogonal Matching Pursuit (OMP)</span></a> by <a class="reference external" href="https://vene.ro/">Vlad Niculae</a>.</p></li>
<li><p><a class="reference internal" href="../modules/decomposition.html#sparsecoder"><span class="std std-ref">Sparse coding with a precomputed dictionary</span></a> by <a class="reference external" href="https://vene.ro/">Vlad Niculae</a>.</p></li>
<li><p><a class="reference internal" href="../modules/clustering.html#mini-batch-kmeans"><span class="std std-ref">Mini Batch K-Means</span></a> performance improvements by <a class="reference external" href="https://twitter.com/ogrisel">Olivier Grisel</a>.</p></li>
<li><p><a class="reference internal" href="../modules/clustering.html#k-means"><span class="std std-ref">K-means</span></a> support for sparse matrices by <a class="reference external" href="http://www.mblondel.org">Mathieu Blondel</a>.</p></li>
<li><p>Improved documentation for developers and for the <a class="reference internal" href="../modules/classes.html#module-sklearn.utils" title="sklearn.utils"><code class="xref py py-mod docutils literal notranslate"><span class="pre">sklearn.utils</span></code></a>
module, by <a class="reference external" href="https://staff.washington.edu/jakevdp/">Jake Vanderplas</a>.</p></li>
<li><p>Vectorized 20newsgroups dataset loader
(<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>) by
<a class="reference external" href="http://www.mblondel.org">Mathieu Blondel</a>.</p></li>
<li><p><a class="reference internal" href="../modules/multiclass.html#multiclass"><span class="std std-ref">Multiclass and multilabel algorithms</span></a> by <a class="reference external" href="https://github.com/larsmans">Lars Buitinck</a>.</p></li>
<li><p>Utilities for fast computation of mean and variance for sparse matrices
by <a class="reference external" href="http://www.mblondel.org">Mathieu Blondel</a>.</p></li>
<li><p>Make <a class="reference internal" href="../modules/generated/sklearn.preprocessing.scale.html#sklearn.preprocessing.scale" title="sklearn.preprocessing.scale"><code class="xref py py-func docutils literal notranslate"><span class="pre">sklearn.preprocessing.scale</span></code></a> and
<code class="xref py py-class docutils literal notranslate"><span class="pre">sklearn.preprocessing.Scaler</span></code> work on sparse matrices by
<a class="reference external" href="https://twitter.com/ogrisel">Olivier Grisel</a></p></li>
<li><p>Feature importances using decision trees and/or forest of trees,
by <a class="reference external" href="http://www.montefiore.ulg.ac.be/~glouppe/">Gilles Louppe</a>.</p></li>
<li><p>Parallel implementation of forests of randomized trees by
<a class="reference external" href="http://www.montefiore.ulg.ac.be/~glouppe/">Gilles Louppe</a>.</p></li>
<li><p><code class="xref py py-class docutils literal notranslate"><span class="pre">sklearn.cross_validation.ShuffleSplit</span></code> can subsample the train
sets as well as the test sets by <a class="reference external" href="https://twitter.com/ogrisel">Olivier Grisel</a>.</p></li>
<li><p>Errors in the build of the documentation fixed by <a class="reference external" href="https://amueller.github.io/">Andreas Müller</a>.</p></li>
</ul>
</div>
<div class="section" id="id7">
<h2>API changes summary<a class="headerlink" href="#id7" title="Permalink to this headline">¶</a></h2>
<p>Here are the code migration instructions when upgrading from scikit-learn
version 0.9:</p>
<ul>
<li><p>Some estimators that may overwrite their inputs to save memory previously
had <code class="docutils literal notranslate"><span class="pre">overwrite_</span></code> parameters; these have been replaced with <code class="docutils literal notranslate"><span class="pre">copy_</span></code>
parameters with exactly the opposite meaning.</p>
<p>This particularly affects some of the estimators in <code class="xref py py-mod docutils literal notranslate"><span class="pre">linear_model</span></code>.
The default behavior is still to copy everything passed in.</p>
</li>
<li><p>The SVMlight dataset loader <a class="reference internal" href="../modules/generated/sklearn.datasets.load_svmlight_file.html#sklearn.datasets.load_svmlight_file" title="sklearn.datasets.load_svmlight_file"><code class="xref py py-func docutils literal notranslate"><span class="pre">sklearn.datasets.load_svmlight_file</span></code></a> no
longer supports loading two files at once; use <code class="docutils literal notranslate"><span class="pre">load_svmlight_files</span></code>
instead. Also, the (unused) <code class="docutils literal notranslate"><span class="pre">buffer_mb</span></code> parameter is gone.</p></li>
<li><p>Sparse estimators in the <a class="reference internal" href="../modules/sgd.html#sgd"><span class="std std-ref">Stochastic Gradient Descent</span></a> module use dense parameter vector
<code class="docutils literal notranslate"><span class="pre">coef_</span></code> instead of <code class="docutils literal notranslate"><span class="pre">sparse_coef_</span></code>. This significantly improves
test time performance.</p></li>
<li><p>The <a class="reference internal" href="../modules/covariance.html#covariance"><span class="std std-ref">Covariance estimation</span></a> module now has a robust estimator of
covariance, the Minimum Covariance Determinant estimator.</p></li>
<li><p>Cluster evaluation metrics in <code class="xref py py-mod docutils literal notranslate"><span class="pre">metrics.cluster</span></code> have been refactored
but the changes are backwards compatible. They have been moved to the
<code class="xref py py-mod docutils literal notranslate"><span class="pre">metrics.cluster.supervised</span></code>, along with
<code class="xref py py-mod docutils literal notranslate"><span class="pre">metrics.cluster.unsupervised</span></code> which contains the Silhouette
Coefficient.</p></li>
<li><p>The <code class="docutils literal notranslate"><span class="pre">permutation_test_score</span></code> function now behaves the same way as
<code class="docutils literal notranslate"><span class="pre">cross_val_score</span></code> (i.e. uses the mean score across the folds.)</p></li>
<li><p>Cross Validation generators now use integer indices (<code class="docutils literal notranslate"><span class="pre">indices=True</span></code>)
by default instead of boolean masks. This make it more intuitive to
use with sparse matrix data.</p></li>
<li><p>The functions used for sparse coding, <code class="docutils literal notranslate"><span class="pre">sparse_encode</span></code> and
<code class="docutils literal notranslate"><span class="pre">sparse_encode_parallel</span></code> have been combined into
<a class="reference internal" href="../modules/generated/sklearn.decomposition.sparse_encode.html#sklearn.decomposition.sparse_encode" title="sklearn.decomposition.sparse_encode"><code class="xref py py-func docutils literal notranslate"><span class="pre">sklearn.decomposition.sparse_encode</span></code></a>, and the shapes of the arrays
have been transposed for consistency with the matrix factorization setting,
as opposed to the regression setting.</p></li>
<li><p>Fixed an off-by-one error in the SVMlight/LibSVM file format handling;
files generated using <a class="reference internal" href="../modules/generated/sklearn.datasets.dump_svmlight_file.html#sklearn.datasets.dump_svmlight_file" title="sklearn.datasets.dump_svmlight_file"><code class="xref py py-func docutils literal notranslate"><span class="pre">sklearn.datasets.dump_svmlight_file</span></code></a> should be
re-generated. (They should continue to work, but accidentally had one
extra column of zeros prepended.)</p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">BaseDictionaryLearning</span></code> class replaced by <code class="docutils literal notranslate"><span class="pre">SparseCodingMixin</span></code>.</p></li>
<li><p><code class="xref py py-func docutils literal notranslate"><span class="pre">sklearn.utils.extmath.fast_svd</span></code> has been renamed
<a class="reference internal" href="../modules/generated/sklearn.utils.extmath.randomized_svd.html#sklearn.utils.extmath.randomized_svd" title="sklearn.utils.extmath.randomized_svd"><code class="xref py py-func docutils literal notranslate"><span class="pre">sklearn.utils.extmath.randomized_svd</span></code></a> and the default
oversampling is now fixed to 10 additional random vectors instead
of doubling the number of components to extract. The new behavior
follows the reference paper.</p></li>
</ul>
</div>
<div class="section" id="id8">
<h2>People<a class="headerlink" href="#id8" title="Permalink to this headline">¶</a></h2>
<p>The following people contributed to scikit-learn since last release:</p>
<blockquote>
<div><ul class="simple">
<li><p>246 <a class="reference external" href="https://amueller.github.io/">Andreas Müller</a></p></li>
<li><p>242 <a class="reference external" href="https://twitter.com/ogrisel">Olivier Grisel</a></p></li>
<li><p>220 <a class="reference external" href="http://www.montefiore.ulg.ac.be/~glouppe/">Gilles Louppe</a></p></li>
<li><p>183 <a class="reference external" href="http://personal.ee.surrey.ac.uk/Personal/B.Holt">Brian Holt</a></p></li>
<li><p>166 <a class="reference external" href="http://gael-varoquaux.info">Gael Varoquaux</a></p></li>
<li><p>144 <a class="reference external" href="https://github.com/larsmans">Lars Buitinck</a></p></li>
<li><p>73 <a class="reference external" href="https://vene.ro/">Vlad Niculae</a></p></li>
<li><p>65 <a class="reference external" href="https://sites.google.com/site/peterprettenhofer/">Peter Prettenhofer</a></p></li>
<li><p>64 <a class="reference external" href="http://fa.bianp.net">Fabian Pedregosa</a></p></li>
<li><p>60 Robert Layton</p></li>
<li><p>55 <a class="reference external" href="http://www.mblondel.org">Mathieu Blondel</a></p></li>
<li><p>52 <a class="reference external" href="https://staff.washington.edu/jakevdp/">Jake Vanderplas</a></p></li>
<li><p>44 Noel Dawe</p></li>
<li><p>38 <a class="reference external" href="http://alexandre.gramfort.net">Alexandre Gramfort</a></p></li>
<li><p>24 <a class="reference external" href="https://github.com/VirgileFritsch">Virgile Fritsch</a></p></li>
<li><p>23 <a class="reference external" href="https://www.mit.edu/~satra/">Satrajit Ghosh</a></p></li>
<li><p>3 Jan Hendrik Metzen</p></li>
<li><p>3 Kenneth C. Arnold</p></li>
<li><p>3 Shiqiao Du</p></li>
<li><p>3 Tim Sheerman-Chase</p></li>
<li><p>3 <a class="reference external" href="http://www.onerussian.com/">Yaroslav Halchenko</a></p></li>
<li><p>2 Bala Subrahmanyam Varanasi</p></li>
<li><p>2 DraXus</p></li>
<li><p>2 Michael Eickenberg</p></li>
<li><p>1 Bogdan Trach</p></li>
<li><p>1 Félix-Antoine Fortin</p></li>
<li><p>1 Juan Manuel Caicedo Carvajal</p></li>
<li><p>1 Nelle Varoquaux</p></li>
<li><p>1 <a class="reference external" href="https://twitter.com/npinto">Nicolas Pinto</a></p></li>
<li><p>1 Tiziano Zito</p></li>
<li><p>1 Xinfan Meng</p></li>
</ul>
</div></blockquote>
</div>
</div>
<div class="section" id="version-0-9">
<span id="changes-0-9"></span><h1>Version 0.9<a class="headerlink" href="#version-0-9" title="Permalink to this headline">¶</a></h1>
<p><strong>September 21, 2011</strong></p>
<p>scikit-learn 0.9 was released on September 2011, three months after the 0.8
release and includes the new modules <a class="reference internal" href="../modules/manifold.html#manifold"><span class="std std-ref">Manifold learning</span></a>, <a class="reference internal" href="../modules/mixture.html#dirichlet-process"><span class="std std-ref">The Dirichlet Process</span></a>
as well as several new algorithms and documentation improvements.</p>
<p>This release also includes the dictionary-learning work developed by
<a class="reference external" href="https://vene.ro/">Vlad Niculae</a> as part of the <a class="reference external" href="https://developers.google.com/open-source/gsoc">Google Summer of Code</a> program.</p>
<p><div style="text-align: center; margin: 0px 0 -5px 0;"> <a class="reference external" href="../auto_examples/linear_model/plot_omp.html"><img alt="banner2" src="../_images/sphx_glr_plot_omp_thumb1.png" /></a> <a class="reference external" href="../auto_examples/manifold/plot_compare_methods.html"><img alt="banner1" src="../_images/sphx_glr_plot_compare_methods_thumb1.png" /></a> <a class="reference external" href="../auto_examples/decomposition/plot_kernel_pca.html"><img alt="banner3" src="../_images/sphx_glr_plot_kernel_pca_thumb1.png" /></a> </div></p>
<div class="section" id="id9">
<h2>Changelog<a class="headerlink" href="#id9" title="Permalink to this headline">¶</a></h2>
<ul class="simple">
<li><p>New <a class="reference internal" href="../modules/manifold.html#manifold"><span class="std std-ref">Manifold learning</span></a> module by <a class="reference external" href="https://staff.washington.edu/jakevdp/">Jake Vanderplas</a> and
<a class="reference external" href="http://fa.bianp.net">Fabian Pedregosa</a>.</p></li>
<li><p>New <a class="reference internal" href="../modules/mixture.html#dirichlet-process"><span class="std std-ref">Dirichlet Process</span></a> Gaussian Mixture
Model by <a class="reference external" href="http://atpassos.me">Alexandre Passos</a></p></li>
<li><p><a class="reference internal" href="../modules/neighbors.html#neighbors"><span class="std std-ref">Nearest Neighbors</span></a> module refactoring by <a class="reference external" href="https://staff.washington.edu/jakevdp/">Jake Vanderplas</a> :
general refactoring, support for sparse matrices in input, speed and
documentation improvements. See the next section for a full list of API
changes.</p></li>
<li><p>Improvements on the <a class="reference internal" href="../modules/feature_selection.html#feature-selection"><span class="std std-ref">Feature selection</span></a> module by
<a class="reference external" href="http://www.montefiore.ulg.ac.be/~glouppe/">Gilles Louppe</a> : refactoring of the RFE classes, documentation
rewrite, increased efficiency and minor API changes.</p></li>
<li><p><a class="reference internal" href="../modules/decomposition.html#sparsepca"><span class="std std-ref">Sparse principal components analysis (SparsePCA and MiniBatchSparsePCA)</span></a> by <a class="reference external" href="https://vene.ro/">Vlad Niculae</a>, <a class="reference external" href="http://gael-varoquaux.info">Gael Varoquaux</a> and
<a class="reference external" href="http://alexandre.gramfort.net">Alexandre Gramfort</a></p></li>
<li><p>Printing an estimator now behaves independently of architectures
and Python version thanks to <a class="reference external" href="https://github.com/JeanKossaifi">Jean Kossaifi</a>.</p></li>
<li><p><a class="reference internal" href="../datasets/index.html#libsvm-loader"><span class="std std-ref">Loader for libsvm/svmlight format</span></a> by
<a class="reference external" href="http://www.mblondel.org">Mathieu Blondel</a> and <a class="reference external" href="https://github.com/larsmans">Lars Buitinck</a></p></li>
<li><p>Documentation improvements: thumbnails in
example gallery by <a class="reference external" href="http://fa.bianp.net">Fabian Pedregosa</a>.</p></li>
<li><p>Important bugfixes in <a class="reference internal" href="../modules/svm.html#svm"><span class="std std-ref">Support Vector Machines</span></a> module (segfaults, bad
performance) by <a class="reference external" href="http://fa.bianp.net">Fabian Pedregosa</a>.</p></li>
<li><p>Added <a class="reference internal" href="../modules/naive_bayes.html#multinomial-naive-bayes"><span class="std std-ref">Multinomial Naive Bayes</span></a> and <a class="reference internal" href="../modules/naive_bayes.html#bernoulli-naive-bayes"><span class="std std-ref">Bernoulli Naive Bayes</span></a>
by <a class="reference external" href="https://github.com/larsmans">Lars Buitinck</a></p></li>
<li><p>Text feature extraction optimizations by Lars Buitinck</p></li>
<li><p>Chi-Square feature selection
(<code class="xref py py-func docutils literal notranslate"><span class="pre">feature_selection.univariate_selection.chi2</span></code>) by <a class="reference external" href="https://github.com/larsmans">Lars Buitinck</a>.</p></li>
<li><p><a class="reference internal" href="../datasets/index.html#sample-generators"><span class="std std-ref">Generated datasets</span></a> module refactoring by <a class="reference external" href="http://www.montefiore.ulg.ac.be/~glouppe/">Gilles Louppe</a></p></li>
<li><p><a class="reference internal" href="../modules/multiclass.html#multiclass"><span class="std std-ref">Multiclass and multilabel algorithms</span></a> by <a class="reference external" href="http://www.mblondel.org">Mathieu Blondel</a></p></li>
<li><p>Ball tree rewrite by <a class="reference external" href="https://staff.washington.edu/jakevdp/">Jake Vanderplas</a></p></li>
<li><p>Implementation of <a class="reference internal" href="../modules/clustering.html#dbscan"><span class="std std-ref">DBSCAN</span></a> algorithm by Robert Layton</p></li>
<li><p>Kmeans predict and transform by Robert Layton</p></li>
<li><p>Preprocessing module refactoring by <a class="reference external" href="https://twitter.com/ogrisel">Olivier Grisel</a></p></li>
<li><p>Faster mean shift by Conrad Lee</p></li>
<li><p>New <code class="docutils literal notranslate"><span class="pre">Bootstrap</span></code>, <a class="reference internal" href="../modules/cross_validation.html#shufflesplit"><span class="std std-ref">Random permutations cross-validation a.k.a. Shuffle & Split</span></a> and various other
improvements in cross validation schemes by <a class="reference external" href="https://twitter.com/ogrisel">Olivier Grisel</a> and
<a class="reference external" href="http://gael-varoquaux.info">Gael Varoquaux</a></p></li>
<li><p>Adjusted Rand index and V-Measure clustering evaluation metrics by <a class="reference external" href="https://twitter.com/ogrisel">Olivier Grisel</a></p></li>
<li><p>Added <a class="reference internal" href="../modules/generated/sklearn.linear_model.OrthogonalMatchingPursuit.html#sklearn.linear_model.OrthogonalMatchingPursuit" title="sklearn.linear_model.OrthogonalMatchingPursuit"><code class="xref py py-class docutils literal notranslate"><span class="pre">Orthogonal</span> <span class="pre">Matching</span> <span class="pre">Pursuit</span></code></a> by <a class="reference external" href="https://vene.ro/">Vlad Niculae</a></p></li>
<li><p>Added 2D-patch extractor utilities in the <a class="reference internal" href="../modules/feature_extraction.html#feature-extraction"><span class="std std-ref">Feature extraction</span></a> module by <a class="reference external" href="https://vene.ro/">Vlad Niculae</a></p></li>
<li><p>Implementation of <a class="reference internal" href="../modules/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">linear_model.LassoLarsCV</span></code></a>
(cross-validated Lasso solver using the Lars algorithm) and
<a class="reference internal" href="../modules/generated/sklearn.linear_model.LassoLarsIC.html#sklearn.linear_model.LassoLarsIC" title="sklearn.linear_model.LassoLarsIC"><code class="xref py py-class docutils literal notranslate"><span class="pre">linear_model.LassoLarsIC</span></code></a> (BIC/AIC model
selection in Lars) by <a class="reference external" href="http://gael-varoquaux.info">Gael Varoquaux</a>
and <a class="reference external" href="http://alexandre.gramfort.net">Alexandre Gramfort</a></p></li>
<li><p>Scalability improvements to <a class="reference internal" href="../modules/generated/sklearn.metrics.roc_curve.html#sklearn.metrics.roc_curve" title="sklearn.metrics.roc_curve"><code class="xref py py-func docutils literal notranslate"><span class="pre">metrics.roc_curve</span></code></a> by Olivier Hervieu</p></li>
<li><p>Distance helper functions <code class="xref py py-func docutils literal notranslate"><span class="pre">metrics.pairwise.pairwise_distances</span></code>
and <a class="reference internal" href="../modules/generated/sklearn.metrics.pairwise.pairwise_kernels.html#sklearn.metrics.pairwise.pairwise_kernels" title="sklearn.metrics.pairwise.pairwise_kernels"><code class="xref py py-func docutils literal notranslate"><span class="pre">metrics.pairwise.pairwise_kernels</span></code></a> by Robert Layton</p></li>
<li><p><a class="reference internal" href="../modules/generated/sklearn.cluster.MiniBatchKMeans.html#sklearn.cluster.MiniBatchKMeans" title="sklearn.cluster.MiniBatchKMeans"><code class="xref py py-class docutils literal notranslate"><span class="pre">Mini-Batch</span> <span class="pre">K-Means</span></code></a> by Nelle Varoquaux and Peter Prettenhofer.</p></li>
<li><p>mldata utilities by Pietro Berkes.</p></li>
<li><p><a class="reference internal" href="../datasets/index.html#olivetti-faces-dataset"><span class="std std-ref">The Olivetti faces dataset</span></a> by <a class="reference external" href="http://www-etud.iro.umontreal.ca/~wardefar/">David Warde-Farley</a>.</p></li>
</ul>
</div>
<div class="section" id="id10">
<h2>API changes summary<a class="headerlink" href="#id10" title="Permalink to this headline">¶</a></h2>
<p>Here are the code migration instructions when upgrading from scikit-learn
version 0.8:</p>
<ul>
<li><p>The <code class="docutils literal notranslate"><span class="pre">scikits.learn</span></code> package was renamed <code class="docutils literal notranslate"><span class="pre">sklearn</span></code>. There is
still a <code class="docutils literal notranslate"><span class="pre">scikits.learn</span></code> package alias for backward compatibility.</p>
<p>Third-party projects with a dependency on scikit-learn 0.9+ should
upgrade their codebase. For instance, under Linux / MacOSX just run
(make a backup first!):</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">find</span> <span class="o">-</span><span class="n">name</span> <span class="s2">"*.py"</span> <span class="o">|</span> <span class="n">xargs</span> <span class="n">sed</span> <span class="o">-</span><span class="n">i</span> <span class="s1">'s/</span><span class="se">\b</span><span class="s1">scikits.learn</span><span class="se">\b</span><span class="s1">/sklearn/g'</span>
</pre></div>
</div>
</li>
<li><p>Estimators no longer accept model parameters as <code class="docutils literal notranslate"><span class="pre">fit</span></code> arguments:
instead all parameters must be only be passed as constructor
arguments or using the now public <code class="docutils literal notranslate"><span class="pre">set_params</span></code> method inherited
from <a class="reference internal" href="../modules/generated/sklearn.base.BaseEstimator.html#sklearn.base.BaseEstimator" title="sklearn.base.BaseEstimator"><code class="xref py py-class docutils literal notranslate"><span class="pre">base.BaseEstimator</span></code></a>.</p>
<p>Some estimators can still accept keyword arguments on the <code class="docutils literal notranslate"><span class="pre">fit</span></code>
but this is restricted to data-dependent values (e.g. a Gram matrix
or an affinity matrix that are precomputed from the <code class="docutils literal notranslate"><span class="pre">X</span></code> data matrix.</p>
</li>
<li><p>The <code class="docutils literal notranslate"><span class="pre">cross_val</span></code> package has been renamed to <code class="docutils literal notranslate"><span class="pre">cross_validation</span></code>
although there is also a <code class="docutils literal notranslate"><span class="pre">cross_val</span></code> package alias in place for
backward compatibility.</p>
<p>Third-party projects with a dependency on scikit-learn 0.9+ should
upgrade their codebase. For instance, under Linux / MacOSX just run
(make a backup first!):</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">find</span> <span class="o">-</span><span class="n">name</span> <span class="s2">"*.py"</span> <span class="o">|</span> <span class="n">xargs</span> <span class="n">sed</span> <span class="o">-</span><span class="n">i</span> <span class="s1">'s/</span><span class="se">\b</span><span class="s1">cross_val</span><span class="se">\b</span><span class="s1">/cross_validation/g'</span>
</pre></div>
</div>
</li>
<li><p>The <code class="docutils literal notranslate"><span class="pre">score_func</span></code> argument of the
<code class="docutils literal notranslate"><span class="pre">sklearn.cross_validation.cross_val_score</span></code> function is now expected
to accept <code class="docutils literal notranslate"><span class="pre">y_test</span></code> and <code class="docutils literal notranslate"><span class="pre">y_predicted</span></code> as only arguments for
classification and regression tasks or <code class="docutils literal notranslate"><span class="pre">X_test</span></code> for unsupervised
estimators.</p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">gamma</span></code> parameter for support vector machine algorithms is set
to <code class="docutils literal notranslate"><span class="pre">1</span> <span class="pre">/</span> <span class="pre">n_features</span></code> by default, instead of <code class="docutils literal notranslate"><span class="pre">1</span> <span class="pre">/</span> <span class="pre">n_samples</span></code>.</p></li>
<li><p>The <code class="docutils literal notranslate"><span class="pre">sklearn.hmm</span></code> has been marked as orphaned: it will be removed
from scikit-learn in version 0.11 unless someone steps up to
contribute documentation, examples and fix lurking numerical
stability issues.</p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">sklearn.neighbors</span></code> has been made into a submodule. The two previously
available estimators, <code class="docutils literal notranslate"><span class="pre">NeighborsClassifier</span></code> and <code class="docutils literal notranslate"><span class="pre">NeighborsRegressor</span></code>
have been marked as deprecated. Their functionality has been divided
among five new classes: <code class="docutils literal notranslate"><span class="pre">NearestNeighbors</span></code> for unsupervised neighbors
searches, <code class="docutils literal notranslate"><span class="pre">KNeighborsClassifier</span></code> & <code class="docutils literal notranslate"><span class="pre">RadiusNeighborsClassifier</span></code>
for supervised classification problems, and <code class="docutils literal notranslate"><span class="pre">KNeighborsRegressor</span></code>
& <code class="docutils literal notranslate"><span class="pre">RadiusNeighborsRegressor</span></code> for supervised regression problems.</p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">sklearn.ball_tree.BallTree</span></code> has been moved to
<code class="docutils literal notranslate"><span class="pre">sklearn.neighbors.BallTree</span></code>. Using the former will generate a warning.</p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">sklearn.linear_model.LARS()</span></code> and related classes (LassoLARS,
LassoLARSCV, etc.) have been renamed to
<code class="docutils literal notranslate"><span class="pre">sklearn.linear_model.Lars()</span></code>.</p></li>
<li><p>All distance metrics and kernels in <code class="docutils literal notranslate"><span class="pre">sklearn.metrics.pairwise</span></code> now have a Y
parameter, which by default is None. If not given, the result is the distance
(or kernel similarity) between each sample in Y. If given, the result is the
pairwise distance (or kernel similarity) between samples in X to Y.</p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">sklearn.metrics.pairwise.l1_distance</span></code> is now called <code class="docutils literal notranslate"><span class="pre">manhattan_distance</span></code>,
and by default returns the pairwise distance. For the component wise distance,
set the parameter <code class="docutils literal notranslate"><span class="pre">sum_over_features</span></code> to <code class="docutils literal notranslate"><span class="pre">False</span></code>.</p></li>
</ul>
<p>Backward compatibility package aliases and other deprecated classes and
functions will be removed in version 0.11.</p>
</div>
<div class="section" id="id11">
<h2>People<a class="headerlink" href="#id11" title="Permalink to this headline">¶</a></h2>
<p>38 people contributed to this release.</p>
<ul class="simple">
<li><p>387 <a class="reference external" href="https://vene.ro/">Vlad Niculae</a></p></li>
<li><p>320 <a class="reference external" href="https://twitter.com/ogrisel">Olivier Grisel</a></p></li>
<li><p>192 <a class="reference external" href="https://github.com/larsmans">Lars Buitinck</a></p></li>
<li><p>179 <a class="reference external" href="http://gael-varoquaux.info">Gael Varoquaux</a></p></li>
<li><p>168 <a class="reference external" href="http://fa.bianp.net">Fabian Pedregosa</a> (<a class="reference external" href="https://www.inria.fr/">INRIA</a>, <a class="reference external" href="http://parietal.saclay.inria.fr/">Parietal Team</a>)</p></li>
<li><p>127 <a class="reference external" href="https://staff.washington.edu/jakevdp/">Jake Vanderplas</a></p></li>
<li><p>120 <a class="reference external" href="http://www.mblondel.org">Mathieu Blondel</a></p></li>
<li><p>85 <a class="reference external" href="http://atpassos.me">Alexandre Passos</a></p></li>
<li><p>67 <a class="reference external" href="http://alexandre.gramfort.net">Alexandre Gramfort</a></p></li>
<li><p>57 <a class="reference external" href="https://sites.google.com/site/peterprettenhofer/">Peter Prettenhofer</a></p></li>
<li><p>56 <a class="reference external" href="http://www.montefiore.ulg.ac.be/~glouppe/">Gilles Louppe</a></p></li>
<li><p>42 Robert Layton</p></li>
<li><p>38 Nelle Varoquaux</p></li>
<li><p>32 <a class="reference external" href="https://github.com/JeanKossaifi">Jean Kossaifi</a></p></li>
<li><p>30 Conrad Lee</p></li>
<li><p>22 Pietro Berkes</p></li>
<li><p>18 andy</p></li>
<li><p>17 David Warde-Farley</p></li>
<li><p>12 Brian Holt</p></li>
<li><p>11 Robert</p></li>
<li><p>8 Amit Aides</p></li>
<li><p>8 <a class="reference external" href="https://github.com/VirgileFritsch">Virgile Fritsch</a></p></li>
<li><p>7 <a class="reference external" href="http://www.onerussian.com/">Yaroslav Halchenko</a></p></li>
<li><p>6 Salvatore Masecchia</p></li>
<li><p>5 Paolo Losi</p></li>
<li><p>4 Vincent Schut</p></li>
<li><p>3 Alexis Metaireau</p></li>
<li><p>3 Bryan Silverthorn</p></li>
<li><p>3 <a class="reference external" href="https://amueller.github.io/">Andreas Müller</a></p></li>
<li><p>2 Minwoo Jake Lee</p></li>
<li><p>1 Emmanuelle Gouillart</p></li>
<li><p>1 Keith Goodman</p></li>
<li><p>1 Lucas Wiman</p></li>
<li><p>1 <a class="reference external" href="https://twitter.com/npinto">Nicolas Pinto</a></p></li>
<li><p>1 Thouis (Ray) Jones</p></li>
<li><p>1 Tim Sheerman-Chase</p></li>
</ul>
</div>
</div>
<div class="section" id="version-0-8">
<span id="changes-0-8"></span><h1>Version 0.8<a class="headerlink" href="#version-0-8" title="Permalink to this headline">¶</a></h1>
<p><strong>May 11, 2011</strong></p>
<p>scikit-learn 0.8 was released on May 2011, one month after the first
“international” <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/wiki/Upcoming-events">scikit-learn coding sprint</a> and is
marked by the inclusion of important modules: <a class="reference internal" href="../modules/clustering.html#hierarchical-clustering"><span class="std std-ref">Hierarchical clustering</span></a>,
<a class="reference internal" href="../modules/cross_decomposition.html#cross-decomposition"><span class="std std-ref">Cross decomposition</span></a>, <a class="reference internal" href="../modules/decomposition.html#nmf"><span class="std std-ref">Non-negative matrix factorization (NMF or NNMF)</span></a>, initial support for Python 3 and by important
enhancements and bug fixes.</p>
<div class="section" id="id12">
<h2>Changelog<a class="headerlink" href="#id12" title="Permalink to this headline">¶</a></h2>
<p>Several new modules where introduced during this release:</p>
<ul class="simple">
<li><p>New <a class="reference internal" href="../modules/clustering.html#hierarchical-clustering"><span class="std std-ref">Hierarchical clustering</span></a> module by Vincent Michel,
<a class="reference external" href="https://team.inria.fr/parietal/bertrand-thirions-page">Bertrand Thirion</a>, <a class="reference external" href="http://alexandre.gramfort.net">Alexandre Gramfort</a> and <a class="reference external" href="http://gael-varoquaux.info">Gael Varoquaux</a>.</p></li>
<li><p><a class="reference internal" href="../modules/decomposition.html#kernel-pca"><span class="std std-ref">Kernel PCA</span></a> implementation by <a class="reference external" href="http://www.mblondel.org">Mathieu Blondel</a></p></li>
<li><p><a class="reference internal" href="../datasets/index.html#labeled-faces-in-the-wild-dataset"><span class="std std-ref">The Labeled Faces in the Wild face recognition dataset</span></a> by <a class="reference external" href="https://twitter.com/ogrisel">Olivier Grisel</a>.</p></li>
<li><p>New <a class="reference internal" href="../modules/cross_decomposition.html#cross-decomposition"><span class="std std-ref">Cross decomposition</span></a> module by <a class="reference external" href="https://sites.google.com/site/duchesnay/home">Edouard Duchesnay</a>.</p></li>
<li><p><a class="reference internal" href="../modules/decomposition.html#nmf"><span class="std std-ref">Non-negative matrix factorization (NMF or NNMF)</span></a> module <a class="reference external" href="https://vene.ro/">Vlad Niculae</a></p></li>
<li><p>Implementation of the <a class="reference internal" href="../modules/covariance.html#oracle-approximating-shrinkage"><span class="std std-ref">Oracle Approximating Shrinkage</span></a> algorithm by
<a class="reference external" href="https://github.com/VirgileFritsch">Virgile Fritsch</a> in the <a class="reference internal" href="../modules/covariance.html#covariance"><span class="std std-ref">Covariance estimation</span></a> module.</p></li>
</ul>
<p>Some other modules benefited from significant improvements or cleanups.</p>
<ul class="simple">
<li><p>Initial support for Python 3: builds and imports cleanly,
some modules are usable while others have failing tests by <a class="reference external" href="http://fa.bianp.net">Fabian Pedregosa</a>.</p></li>
<li><p><a class="reference internal" href="../modules/generated/sklearn.decomposition.PCA.html#sklearn.decomposition.PCA" title="sklearn.decomposition.PCA"><code class="xref py py-class docutils literal notranslate"><span class="pre">decomposition.PCA</span></code></a> is now usable from the Pipeline object by <a class="reference external" href="https://twitter.com/ogrisel">Olivier Grisel</a>.</p></li>
<li><p>Guide <a class="reference internal" href="../developers/performance.html#performance-howto"><span class="std std-ref">How to optimize for speed</span></a> by <a class="reference external" href="https://twitter.com/ogrisel">Olivier Grisel</a>.</p></li>
<li><p>Fixes for memory leaks in libsvm bindings, 64-bit safer BallTree by Lars Buitinck.</p></li>
<li><p>bug and style fixing in <a class="reference internal" href="../modules/clustering.html#k-means"><span class="std std-ref">K-means</span></a> algorithm by Jan Schlüter.</p></li>
<li><p>Add attribute converged to Gaussian Mixture Models by Vincent Schut.</p></li>
<li><p>Implemented <code class="docutils literal notranslate"><span class="pre">transform</span></code>, <code class="docutils literal notranslate"><span class="pre">predict_log_proba</span></code> in
<a class="reference internal" href="../modules/generated/sklearn.discriminant_analysis.LinearDiscriminantAnalysis.html#sklearn.discriminant_analysis.LinearDiscriminantAnalysis" title="sklearn.discriminant_analysis.LinearDiscriminantAnalysis"><code class="xref py py-class docutils literal notranslate"><span class="pre">discriminant_analysis.LinearDiscriminantAnalysis</span></code></a> By <a class="reference external" href="http://www.mblondel.org">Mathieu Blondel</a>.</p></li>
<li><p>Refactoring in the <a class="reference internal" href="../modules/svm.html#svm"><span class="std std-ref">Support Vector Machines</span></a> module and bug fixes by <a class="reference external" href="http://fa.bianp.net">Fabian Pedregosa</a>,
<a class="reference external" href="http://gael-varoquaux.info">Gael Varoquaux</a> and Amit Aides.</p></li>
<li><p>Refactored SGD module (removed code duplication, better variable naming),
added interface for sample weight by <a class="reference external" href="https://sites.google.com/site/peterprettenhofer/">Peter Prettenhofer</a>.</p></li>
<li><p>Wrapped BallTree with Cython by Thouis (Ray) Jones.</p></li>
<li><p>Added function <a class="reference internal" href="../modules/generated/sklearn.svm.l1_min_c.html#sklearn.svm.l1_min_c" title="sklearn.svm.l1_min_c"><code class="xref py py-func docutils literal notranslate"><span class="pre">svm.l1_min_c</span></code></a> by Paolo Losi.</p></li>
<li><p>Typos, doc style, etc. by <a class="reference external" href="http://www.onerussian.com/">Yaroslav Halchenko</a>, <a class="reference external" href="http://gael-varoquaux.info">Gael Varoquaux</a>,
<a class="reference external" href="https://twitter.com/ogrisel">Olivier Grisel</a>, Yann Malet, <a class="reference external" href="https://twitter.com/npinto">Nicolas Pinto</a>, Lars Buitinck and
<a class="reference external" href="http://fa.bianp.net">Fabian Pedregosa</a>.</p></li>
</ul>
</div>
<div class="section" id="id13">
<h2>People<a class="headerlink" href="#id13" title="Permalink to this headline">¶</a></h2>
<p>People that made this release possible preceded by number of commits:</p>
<ul class="simple">
<li><p>159 <a class="reference external" href="https://twitter.com/ogrisel">Olivier Grisel</a></p></li>
<li><p>96 <a class="reference external" href="http://gael-varoquaux.info">Gael Varoquaux</a></p></li>
<li><p>96 <a class="reference external" href="https://vene.ro/">Vlad Niculae</a></p></li>
<li><p>94 <a class="reference external" href="http://fa.bianp.net">Fabian Pedregosa</a></p></li>
<li><p>36 <a class="reference external" href="http://alexandre.gramfort.net">Alexandre Gramfort</a></p></li>
<li><p>32 Paolo Losi</p></li>
<li><p>31 <a class="reference external" href="https://sites.google.com/site/duchesnay/home">Edouard Duchesnay</a></p></li>
<li><p>30 <a class="reference external" href="http://www.mblondel.org">Mathieu Blondel</a></p></li>
<li><p>25 <a class="reference external" href="https://sites.google.com/site/peterprettenhofer/">Peter Prettenhofer</a></p></li>
<li><p>22 <a class="reference external" href="https://twitter.com/npinto">Nicolas Pinto</a></p></li>
<li><dl class="simple">
<dt>11 <a class="reference external" href="https://github.com/VirgileFritsch">Virgile Fritsch</a></dt><dd><ul>
<li><p>7 Lars Buitinck</p></li>
<li><p>6 Vincent Michel</p></li>