forked from jaquesgrobler/Online-Scikit-Learn-stat-tut
-
Notifications
You must be signed in to change notification settings - Fork 0
/
whats_new.html
1002 lines (975 loc) · 80.7 KB
/
whats_new.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 PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
"http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
<html xmlns="http://www.w3.org/1999/xhtml">
<head>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
<title>0.11 — scikit-learn 0.11-git documentation</title>
<link rel="stylesheet" href="_static/nature.css" type="text/css" />
<link rel="stylesheet" href="_static/pygments.css" type="text/css" />
<script type="text/javascript">
var DOCUMENTATION_OPTIONS = {
URL_ROOT: '',
VERSION: '0.11-git',
COLLAPSE_INDEX: false,
FILE_SUFFIX: '.html',
HAS_SOURCE: true
};
</script>
<script type="text/javascript" src="_static/jquery.js"></script>
<script type="text/javascript" src="_static/underscore.js"></script>
<script type="text/javascript" src="_static/doctools.js"></script>
<link rel="shortcut icon" href="_static/favicon.ico"/>
<link rel="author" title="About these documents" href="about.html" />
<link rel="top" title="scikit-learn 0.11-git documentation" href="index.html" />
<link rel="prev" title="Support" href="support.html" />
<script type="text/javascript">
var _gaq = _gaq || [];
_gaq.push(['_setAccount', 'UA-22606712-2']);
_gaq.push(['_trackPageview']);
(function() {
var ga = document.createElement('script'); ga.type = 'text/javascript'; ga.async = true;
ga.src = ('https:' == document.location.protocol ? 'https://ssl' : 'http://www') + '.google-analytics.com/ga.js';
var s = document.getElementsByTagName('script')[0]; s.parentNode.insertBefore(ga, s);
})();
</script>
</head>
<body>
<div class="header-wrapper">
<div class="header">
<p class="logo"><a href="index.html">
<img src="_static/scikit-learn-logo-small.png" alt="Logo"/>
</a>
</p><div class="navbar">
<ul>
<li><a href="install.html">Download</a></li>
<li><a href="support.html">Support</a></li>
<li><a href="user_guide.html">User Guide</a></li>
<li><a href="auto_examples/index.html">Examples</a></li>
<li><a href="modules/classes.html">Reference</a></li>
</ul>
<div class="search_form">
<div id="cse" style="width: 100%;"></div>
<script src="http://www.google.com/jsapi" type="text/javascript"></script>
<script type="text/javascript">
google.load('search', '1', {language : 'en'});
google.setOnLoadCallback(function() {
var customSearchControl = new google.search.CustomSearchControl('016639176250731907682:tjtqbvtvij0');
customSearchControl.setResultSetSize(google.search.Search.FILTERED_CSE_RESULTSET);
var options = new google.search.DrawOptions();
options.setAutoComplete(true);
customSearchControl.draw('cse', options);
}, true);
</script>
</div>
</div> <!-- end navbar --></div>
</div>
<div class="content-wrapper">
<div class="sphinxsidebar">
<div class="rel rellarge">
<!-- rellinks[1:] is an ugly hack to avoid link to module
index -->
<div class="rellink">
<a href="support.html" title="Support"
accesskey="P">Previous
<br>
<span class="smallrellink">
Support
</span>
<span class="hiddenrellink">
Support
</span>
</a>
</div>
<!-- Ad a link to the 'up' page -->
</div>
<p style="text-align: center">This documentation is
for scikit-learn <strong>version 0.11-git</strong>
— <a href="http://scikit-learn.org/stable/support.html#documentation-resources">Other versions</a></p>
<h3>Citing</h3>
<p>If you use the software, please consider
<a href="about.html#citing-scikit-learn">citing scikit-learn</a>.</p>
<h3>This page</h3>
<ul>
<li><a class="reference internal" href="#">0.11</a><ul>
<li><a class="reference internal" href="#changelog">Changelog</a></li>
<li><a class="reference internal" href="#api-changes-summary">API changes summary</a></li>
</ul>
</li>
<li><a class="reference internal" href="#changes-0-10">0.10</a><ul>
<li><a class="reference internal" href="#id3">Changelog</a></li>
<li><a class="reference internal" href="#id4">API changes summary</a></li>
<li><a class="reference internal" href="#people">People</a></li>
</ul>
</li>
<li><a class="reference internal" href="#changes-0-9">0.9</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="#changes-0-8">0.8</a><ul>
<li><a class="reference internal" href="#id10">Changelog</a></li>
<li><a class="reference internal" href="#id11">People</a></li>
</ul>
</li>
<li><a class="reference internal" href="#changes-0-7">0.7</a><ul>
<li><a class="reference internal" href="#id13">Changelog</a></li>
<li><a class="reference internal" href="#id14">People</a></li>
</ul>
</li>
<li><a class="reference internal" href="#changes-0-6">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="#changes-0-5">0.5</a><ul>
<li><a class="reference internal" href="#id19">Changelog</a><ul>
<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>
</ul>
</li>
<li><a class="reference internal" href="#authors">Authors</a></li>
</ul>
</li>
<li><a class="reference internal" href="#id20">0.4</a><ul>
<li><a class="reference internal" href="#id21">Changelog</a></li>
<li><a class="reference internal" href="#id22">Authors</a></li>
</ul>
</li>
</ul>
</div>
<div class="content">
<div class="documentwrapper">
<div class="bodywrapper">
<div class="body">
<div class="section" id="changes-0-11">
<span id="id1"></span><h1>0.11<a class="headerlink" href="#changes-0-11" title="Permalink to this headline">¶</a></h1>
<div class="section" id="changelog">
<h2>Changelog<a class="headerlink" href="#changelog" title="Permalink to this headline">¶</a></h2>
<blockquote>
<div><ul class="simple">
<li>Merged dense and sparse implementations and added <cite>partial_fit</cite> (support
for online/minibatch learning) and warm_start to the <a class="reference internal" href="modules/sgd.html#sgd"><em>Stochastic Gradient Descent</em></a> module by
<a class="reference external" href="http://www.mblondel.org/journal/">Mathieu Blondel</a>.</li>
<li>Dense and sparse implementations of <a class="reference internal" href="modules/svm.html#svm"><em>Support Vector Machines</em></a> classes and
<a class="reference internal" href="modules/generated/sklearn.linear_model.LogisticRegression.html#sklearn.linear_model.LogisticRegression" title="sklearn.linear_model.LogisticRegression"><tt class="xref py py-class docutils literal"><span class="pre">linear_model.LogisticRegression</span></tt></a> merged by <a class="reference external" href="https://github.com/larsmans">Lars Buitinck</a>.</li>
<li>Regressors can now be used as base estimator in the <a class="reference internal" href="modules/multiclass.html#multiclass"><em>Multiclass and multilabel algorithms</em></a>
module by <a class="reference external" href="http://www.mblondel.org/journal/">Mathieu Blondel</a>.</li>
<li>Added Matthews correlation coefficient (<tt class="xref py py-func docutils literal"><span class="pre">metrics.matthews_corrcoef</span></tt>)
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"><tt class="xref py py-func docutils literal"><span class="pre">metrics.precision_score</span></tt></a>, <a class="reference internal" href="modules/generated/sklearn.metrics.recall_score.html#sklearn.metrics.recall_score" title="sklearn.metrics.recall_score"><tt class="xref py py-func docutils literal"><span class="pre">metrics.recall_score</span></tt></a> and
<a class="reference internal" href="modules/generated/sklearn.metrics.f1_score.html#sklearn.metrics.f1_score" title="sklearn.metrics.f1_score"><tt class="xref py py-func docutils literal"><span class="pre">metrics.f1_score</span></tt></a> by <a class="reference external" href="http://www.mit.edu/~satra/">Satrajit Ghosh</a>.</li>
<li>Added n_jobs option to <a class="reference internal" href="modules/generated/sklearn.metrics.pairwise.pairwise_distances.html#sklearn.metrics.pairwise.pairwise_distances" title="sklearn.metrics.pairwise.pairwise_distances"><tt class="xref py py-func docutils literal"><span class="pre">metrics.pairwise.pairwise_distances</span></tt></a>
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"><tt class="xref py py-func docutils literal"><span class="pre">metrics.pairwise.pairwise_kernels</span></tt></a> for parallel computation,
by <a class="reference external" href="http://www.mblondel.org/journal/">Mathieu Blondel</a>.</li>
<li><a class="reference internal" href="modules/grid_search.html#out-of-bag"><em>Out of Bag Estimates</em></a> of generalization error for <a class="reference internal" href="modules/ensemble.html#ensemble"><em>Ensemble methods</em></a>
by <a class="reference external" href="http://www.ais.uni-bonn.de/~amueller/">Andreas Müller</a>.</li>
<li><a class="reference internal" href="modules/feature_selection.html#randomized-l1"><em>Randomized sparse models</em></a>: Randomized sparse linear models for feature
selection, by <a class="reference external" href="http://www-sop.inria.fr/members/Alexandre.Gramfort/">Alexandre Gramfort</a> and <a class="reference external" href="http://gael-varoquaux.info">Gael Varoquaux</a></li>
<li><a class="reference internal" href="modules/label_propagation.html#label-propagation"><em>Label Propagation</em></a> for semi-supervised learning, by Clay
Woolam. <strong>Note</strong> the semi-supervised API is still work in progress,
and may change.</li>
<li>Added BIC/AIC model selection to classical <a class="reference internal" href="modules/mixture.html#gmm"><em>Gaussian mixture models</em></a> and unified
the API with the remainder of scikit-learn, by <a class="reference external" href="http://parietal.saclay.inria.fr/Members/bertrand-thirion">Bertrand Thirion</a></li>
<li><a class="reference internal" href="modules/clustering.html#k-means"><em>K-means</em></a> can now be run in parallel, using the <cite>n_jobs</cite> argument
to either <a class="reference internal" href="modules/clustering.html#k-means"><em>K-means</em></a> or <tt class="xref py py-class docutils literal"><span class="pre">KMeans</span></tt>, by <a class="reference external" href="http://www.twitter.com/robertlayton">Robert Layton</a>.</li>
<li>Improved <a class="reference internal" href="modules/cross_validation.html#cross-validation"><em>Cross-Validation</em></a> and <a class="reference internal" href="modules/grid_search.html#grid-search"><em>Grid Search</em></a> documentation
and introduced the new <a class="reference internal" href="modules/generated/sklearn.cross_validation.train_test_split.html#sklearn.cross_validation.train_test_split" title="sklearn.cross_validation.train_test_split"><tt class="xref py py-func docutils literal"><span class="pre">cross_validation.train_test_split</span></tt></a>
helper function by <a class="reference external" href="http://twitter.com/ogrisel">Olivier Grisel</a></li>
<li><a class="reference internal" href="modules/generated/sklearn.svm.SVC.html#sklearn.svm.SVC" title="sklearn.svm.SVC"><tt class="xref py py-class docutils literal"><span class="pre">svm.SVC</span></tt></a> members <cite>coef_</cite> and <cite>intercept_</cite> changed sign for consistency
with <cite>decision_function</cite>; for <tt class="docutils literal"><span class="pre">kernel==linear</span></tt>, <cite>coef_</cite> was fixed
in the the one-vs-one case, by <a class="reference external" href="http://www.ais.uni-bonn.de/~amueller/">Andreas Müller</a>.</li>
<li>Performance improvements to efficient leave-one-out cross-validated
Ridge regression, esp. for the <tt class="docutils literal"><span class="pre">n_samples</span> <span class="pre">></span> <span class="pre">n_features</span></tt> case, in
<a class="reference internal" href="modules/generated/sklearn.linear_model.RidgeCV.html#sklearn.linear_model.RidgeCV" title="sklearn.linear_model.RidgeCV"><tt class="xref py py-class docutils literal"><span class="pre">linear_model.RidgeCV</span></tt></a>, by Reuben Fletcher-Costin.</li>
</ul>
</div></blockquote>
</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>
<blockquote>
<div><ul class="simple">
<li><cite>NeighborsClassifier</cite> and <cite>NeighborsRegressor</cite> are gone in the module
<a class="reference internal" href="modules/neighbors.html#neighbors"><em>Nearest Neighbors</em></a>. Use the classes <tt class="xref py py-class docutils literal"><span class="pre">KNeighborsClassifier</span></tt>,
<tt class="xref py py-class docutils literal"><span class="pre">RadiusNeighborsClassifier</span></tt>, <tt class="xref py py-class docutils literal"><span class="pre">KNeighborsRegressor</span></tt>
and/or <tt class="xref py py-class docutils literal"><span class="pre">RadiusNeighborsRegressor</span></tt> instead.</li>
<li>Sparse classes in the <a class="reference internal" href="modules/sgd.html#sgd"><em>Stochastic Gradient Descent</em></a> module are now deprecated.</li>
<li>methods <cite>rvs</cite> and <cite>decode</cite> in <tt class="xref py py-class docutils literal"><span class="pre">GMM</span></tt> module are now deprecated.
<cite>sample</cite> and <cite>score</cite> or <cite>predict</cite> should be used instead.</li>
<li>attribute <cite>_scores</cite> and <cite>_pvalues</cite> in univariate feature selection
objects are now deprecated.
<cite>scores_</cite> or <cite>pvalues_</cite> should be used instead.</li>
<li>In <tt class="xref py py-class docutils literal"><span class="pre">LogisticRegression</span></tt>, <tt class="xref py py-class docutils literal"><span class="pre">LinearSVC</span></tt>, <tt class="xref py py-class docutils literal"><span class="pre">SVC</span></tt> and
<tt class="xref py py-class docutils literal"><span class="pre">NuSVC</span></tt>, the <cite>class_weight</cite> parameter is now an initialization
parameter, not a parameter to fit. This makes grid searches
over this parameter possible.</li>
<li>LFW <tt class="docutils literal"><span class="pre">data</span></tt> is now always shape <tt class="docutils literal"><span class="pre">(n_samples,</span> <span class="pre">n_features)</span></tt> to be
consistent with the Olivetti faces dataset. Use <tt class="docutils literal"><span class="pre">images</span></tt> and
<tt class="docutils literal"><span class="pre">pairs</span></tt> attribute to access the natural images shapes instead.</li>
<li>Setting scale_C=True by default in SVM and LogisticRegression
models. This allows to have a regularization parameter independent
of the number of samples. The scale_C parameter will disappear in v0.12.</li>
<li>In <a class="reference internal" href="modules/generated/sklearn.svm.LinearSVC.html#sklearn.svm.LinearSVC" title="sklearn.svm.LinearSVC"><tt class="xref py py-class docutils literal"><span class="pre">svm.LinearSVC</span></tt></a>, the meaning of the <cite>multi_class</cite> parameter changed.
Options now are ‘ovr’ and ‘crammer_singer’, with ‘ovr’ being the default.
This does not change the default behavior but hopefully is less confusing.</li>
</ul>
</div></blockquote>
</div>
</div>
<div class="section" id="changes-0-10">
<span id="id2"></span><h1>0.10<a class="headerlink" href="#changes-0-10" title="Permalink to this headline">¶</a></h1>
<div class="section" id="id3">
<h2>Changelog<a class="headerlink" href="#id3" title="Permalink to this headline">¶</a></h2>
<blockquote>
<div><ul class="simple">
<li>Python 2.5 compatibility was dropped; the minimum Python version needed
to use scikit-learn is now 2.6.</li>
<li><a class="reference internal" href="modules/covariance.html#sparse-inverse-covariance"><em>Sparse inverse covariance</em></a> estimation using the graph Lasso, with
associated cross-validated estimator, by <a class="reference external" href="http://gael-varoquaux.info">Gael Varoquaux</a></li>
<li>New <a class="reference internal" href="modules/tree.html#tree"><em>Tree</em></a> module by <a class="reference external" href="http://info.ee.surrey.ac.uk/Personal/B.Holt/">Brian Holt</a>, <a class="reference external" href="http://sites.google.com/site/peterprettenhofer/">Peter Prettenhofer</a>,
<a class="reference external" href="http://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.</li>
<li>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).</li>
<li>Fixed a memory leak in in <a class="reference internal" href="modules/svm.html#svm"><em>Support Vector Machines</em></a> module by <a class="reference external" href="http://info.ee.surrey.ac.uk/Personal/B.Holt/">Brian Holt</a> (issue #367).</li>
<li>Faster tests by <a class="reference external" href="http://fseoane.net/blog/">Fabian Pedregosa</a> and others.</li>
<li>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"><tt class="xref py py-func docutils literal"><span class="pre">sklearn.metrics.silhouette_score</span></tt></a> by Robert Layton.</li>
<li>Fixed a bug in <a class="reference internal" href="modules/clustering.html#k-means"><em>K-means</em></a> in the handling of the <tt class="docutils literal"><span class="pre">n_init</span></tt> parameter:
the clustering algorithm used to be run <tt class="docutils literal"><span class="pre">n_init</span></tt> times but the last
solution was retained instead of the best solution by <a class="reference external" href="http://twitter.com/ogrisel">Olivier Grisel</a>.</li>
<li>Minor refactoring in <a class="reference internal" href="modules/sgd.html#sgd"><em>Stochastic Gradient Descent</em></a> module; consolidated dense and sparse
predict methods; Enhanced test time performance by converting model
paramters to fortran-style arrays after fitting (only multi-class).</li>
<li>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"><tt class="xref py py-func docutils literal"><span class="pre">sklearn.metrics.adjusted_mutual_info_score</span></tt></a> by Robert Layton.</li>
<li>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://www-sop.inria.fr/members/Alexandre.Gramfort/">Alexandre Gramfort</a>.</li>
<li>New <a class="reference internal" href="modules/ensemble.html#ensemble"><em>Ensemble Methods</em></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://info.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.</li>
<li><a class="reference internal" href="modules/outlier_detection.html#outlier-detection"><em>Novelty and Outlier Detection</em></a>: outlier and novelty detection, by
<a class="reference external" href="http://parietal.saclay.inria.fr/Members/virgile-fritsch">Virgile Fritsch</a>.</li>
<li><a class="reference internal" href="modules/kernel_approximation.html#kernel-approximation"><em>Kernel Approximation</em></a>: a transform implementing kernel
approximation for fast SGD on non-linear kernels by
<a class="reference external" href="http://www.ais.uni-bonn.de/~amueller/">Andreas Müller</a>.</li>
<li>Fixed a bug due to atom swapping in <a class="reference internal" href="modules/linear_model.html#omp"><em>Orthogonal Matching Pursuit (OMP)</em></a> by <a class="reference external" href="http://vene.ro">Vlad Niculae</a>.</li>
<li><a class="reference internal" href="modules/decomposition.html#sparsecoder"><em>Sparse coding with a precomputed dictionary</em></a> by <a class="reference external" href="http://vene.ro">Vlad Niculae</a>.</li>
<li><a class="reference internal" href="modules/clustering.html#mini-batch-kmeans"><em>Mini Batch K-Means</em></a> performance improvements by <a class="reference external" href="http://twitter.com/ogrisel">Olivier Grisel</a>.</li>
<li><a class="reference internal" href="modules/clustering.html#k-means"><em>K-means</em></a> support for sparse matrices by <a class="reference external" href="http://www.mblondel.org/journal/">Mathieu Blondel</a>.</li>
<li>Improved documentation for developers and for the <a class="reference internal" href="modules/classes.html#module-sklearn.utils" title="sklearn.utils"><tt class="xref py py-mod docutils literal"><span class="pre">sklearn.utils</span></tt></a>
module, by <a class="reference external" href="http://www.astro.washington.edu/users/vanderplas/">Jake VanderPlas</a>.</li>
<li>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"><tt class="xref py py-func docutils literal"><span class="pre">sklearn.datasets.fetch_20newsgroups_vectorized</span></tt></a>) by
<a class="reference external" href="http://www.mblondel.org/journal/">Mathieu Blondel</a>.</li>
<li><a class="reference internal" href="modules/multiclass.html#multiclass"><em>Multiclass and multilabel algorithms</em></a> by <a class="reference external" href="https://github.com/larsmans">Lars Buitinck</a>.</li>
<li>Utilities for fast computation of mean and variance for sparse matrices
by <a class="reference external" href="http://www.mblondel.org/journal/">Mathieu Blondel</a>.</li>
<li>Make <a class="reference internal" href="modules/generated/sklearn.preprocessing.scale.html#sklearn.preprocessing.scale" title="sklearn.preprocessing.scale"><tt class="xref py py-func docutils literal"><span class="pre">sklearn.preprocessing.scale</span></tt></a> and
<a class="reference internal" href="modules/generated/sklearn.preprocessing.Scaler.html#sklearn.preprocessing.Scaler" title="sklearn.preprocessing.Scaler"><tt class="xref py py-class docutils literal"><span class="pre">sklearn.preprocessing.Scaler</span></tt></a> work on sparse matrices by
<a class="reference external" href="http://twitter.com/ogrisel">Olivier Grisel</a></li>
<li>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>.</li>
<li>Parallel implementation of forests of randomized trees by
<a class="reference external" href="http://www.montefiore.ulg.ac.be/~glouppe/">Gilles Louppe</a>.</li>
<li><a class="reference internal" href="modules/generated/sklearn.cross_validation.ShuffleSplit.html#sklearn.cross_validation.ShuffleSplit" title="sklearn.cross_validation.ShuffleSplit"><tt class="xref py py-class docutils literal"><span class="pre">sklearn.cross_validation.ShuffleSplit</span></tt></a> can subsample the train
sets as well as the test sets by <a class="reference external" href="http://twitter.com/ogrisel">Olivier Grisel</a>.</li>
<li>Errors in the build of the documentation fixed by <a class="reference external" href="http://www.ais.uni-bonn.de/~amueller/">Andreas Müller</a>.</li>
</ul>
</div></blockquote>
</div>
<div class="section" id="id4">
<h2>API changes summary<a class="headerlink" href="#id4" title="Permalink to this headline">¶</a></h2>
<p>Here are the code migration instructions when updgrading from scikit-learn
version 0.9:</p>
<blockquote>
<div><ul>
<li><p class="first">Some estimators that may overwrite their inputs to save memory previously
had <tt class="docutils literal"><span class="pre">overwrite_</span></tt> parameters; these have been replaced with <tt class="docutils literal"><span class="pre">copy_</span></tt>
parameters with exactly the opposite meaning.</p>
<p>This particularly affects some of the estimators in <tt class="xref py py-mod docutils literal"><span class="pre">linear_model</span></tt>.
The default behavior is still to copy everything passed in.</p>
</li>
<li><p class="first">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"><tt class="xref py py-func docutils literal"><span class="pre">sklearn.datasets.load_svmlight_file</span></tt></a> no
longer supports loading two files at once; use <tt class="docutils literal"><span class="pre">load_svmlight_files</span></tt>
instead. Also, the (unused) <tt class="docutils literal"><span class="pre">buffer_mb</span></tt> parameter is gone.</p>
</li>
<li><p class="first">Sparse estimators in the <a class="reference internal" href="modules/sgd.html#sgd"><em>Stochastic Gradient Descent</em></a> module use dense parameter vector
<tt class="docutils literal"><span class="pre">coef_</span></tt> instead of <tt class="docutils literal"><span class="pre">sparse_coef_</span></tt>. This significantly improves
test time performance.</p>
</li>
<li><p class="first">The <a class="reference internal" href="modules/covariance.html#covariance"><em>Covariance estimation</em></a> module now has a robust estimator of
covariance, the Minimum Covariance Determinant estimator.</p>
</li>
<li><p class="first">Cluster evaluation metrics in <tt class="xref py py-mod docutils literal"><span class="pre">metrics.cluster</span></tt> have been refactored
but the changes are backwards compatible. They have been moved to the
<tt class="xref py py-mod docutils literal"><span class="pre">metrics.cluster.supervised</span></tt>, along with
<tt class="xref py py-mod docutils literal"><span class="pre">metrics.cluster.unsupervised</span></tt> which contains the Silhouette
Coefficient.</p>
</li>
<li><p class="first">The <tt class="docutils literal"><span class="pre">permutation_test_score</span></tt> function now behaves the same way as
<tt class="docutils literal"><span class="pre">cross_val_score</span></tt> (i.e. uses the mean score across the folds.)</p>
</li>
<li><p class="first">Cross Validation generators now use integer indices (<tt class="docutils literal"><span class="pre">indices=True</span></tt>)
by default instead of boolean masks. This make it more intuitive to
use with sparse matrix data.</p>
</li>
<li><p class="first">The functions used for sparse coding, <tt class="docutils literal"><span class="pre">sparse_encode</span></tt> and
<tt class="docutils literal"><span class="pre">sparse_encode_parallel</span></tt> 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"><tt class="xref py py-func docutils literal"><span class="pre">sklearn.decomposition.sparse_encode</span></tt></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 class="first">Fixed an off-by-one error in the SVMlight/LibSVM file format handling;
files generated using <tt class="xref py py-func docutils literal"><span class="pre">sklearn.datasets.dump_svmlight_file</span></tt> should be
re-generated. (They should continue to work, but accidentally had one
extra column of zeros prepended.)</p>
</li>
<li><p class="first"><tt class="docutils literal"><span class="pre">BaseDictionaryLearning</span></tt> class replaced by <tt class="docutils literal"><span class="pre">SparseCodingMixin</span></tt>.</p>
</li>
<li><p class="first"><tt class="xref py py-func docutils literal"><span class="pre">sklearn.utils.extmath.fast_svd</span></tt> has been renamed
<tt class="xref py py-func docutils literal"><span class="pre">sklearn.utils.extmath.randomized_svd</span></tt> 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></blockquote>
</div>
<div class="section" id="people">
<h2>People<a class="headerlink" href="#people" 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>246 <a class="reference external" href="http://www.ais.uni-bonn.de/~amueller/">Andreas Müller</a></li>
<li>242 <a class="reference external" href="http://twitter.com/ogrisel">Olivier Grisel</a></li>
<li>220 <a class="reference external" href="http://www.montefiore.ulg.ac.be/~glouppe/">Gilles Louppe</a></li>
<li>183 <a class="reference external" href="http://info.ee.surrey.ac.uk/Personal/B.Holt/">Brian Holt</a></li>
<li>166 <a class="reference external" href="http://gael-varoquaux.info">Gael Varoquaux</a></li>
<li>144 <a class="reference external" href="https://github.com/larsmans">Lars Buitinck</a></li>
<li>73 <a class="reference external" href="http://vene.ro">Vlad Niculae</a></li>
<li>65 <a class="reference external" href="http://sites.google.com/site/peterprettenhofer/">Peter Prettenhofer</a></li>
<li>64 <a class="reference external" href="http://fseoane.net/blog/">Fabian Pedregosa</a></li>
<li>60 Robert Layton</li>
<li>55 <a class="reference external" href="http://www.mblondel.org/journal/">Mathieu Blondel</a></li>
<li>52 <a class="reference external" href="http://www.astro.washington.edu/users/vanderplas/">Jake Vanderplas</a></li>
<li>44 Noel Dawe</li>
<li>38 <a class="reference external" href="http://www-sop.inria.fr/members/Alexandre.Gramfort/">Alexandre Gramfort</a></li>
<li>24 <a class="reference external" href="http://parietal.saclay.inria.fr/Members/virgile-fritsch">Virgile Fritsch</a></li>
<li>23 <a class="reference external" href="http://www.mit.edu/~satra/">Satrajit Ghosh</a></li>
<li>3 Jan Hendrik Metzen</li>
<li>3 Kenneth C. Arnold</li>
<li>3 Shiqiao Du</li>
<li>3 Tim Sheerman-Chase</li>
<li>3 <a class="reference external" href="http://www.onerussian.com/">Yaroslav Halchenko</a></li>
<li>2 Bala Subrahmanyam Varanasi</li>
<li>2 DraXus</li>
<li>2 Michael Eickenberg</li>
<li>1 Bogdan Trach</li>
<li>1 Félix-Antoine Fortin</li>
<li>1 Juan Manuel Caicedo Carvajal</li>
<li>1 Nelle Varoquaux</li>
<li>1 <a class="reference external" href="http://pinto.scripts.mit.edu/">Nicolas Pinto</a></li>
<li>1 Tiziano Zito</li>
<li>1 Xinfan Meng</li>
</ul>
</div></blockquote>
</div>
</div>
<div class="section" id="changes-0-9">
<span id="id5"></span><h1>0.9<a class="headerlink" href="#changes-0-9" title="Permalink to this headline">¶</a></h1>
<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"><em>Manifold learning</em></a>, <a class="reference internal" href="modules/mixture.html#dirichlet-process"><em>The Dirichlet Process</em></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="http://vene.ro">Vlad Niculae</a> as part of the <a class="reference external" href="http://code.google.com/soc/">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/plot_omp1.png" /></a> <a class="reference external" href="auto_examples/manifold/plot_compare_methods.html"><img alt="banner1" src="_images/plot_compare_methods1.png" /></a> <a class="reference external" href="auto_examples/decomposition/plot_kernel_pca.html"><img alt="banner3" src="_images/plot_kernel_pca1.png" /></a> </div></p>
<div class="section" id="id6">
<h2>Changelog<a class="headerlink" href="#id6" title="Permalink to this headline">¶</a></h2>
<blockquote>
<div><ul class="simple">
<li>New <a class="reference internal" href="modules/manifold.html#manifold"><em>Manifold learning</em></a> module by <a class="reference external" href="http://www.astro.washington.edu/users/vanderplas/">Jake Vanderplas</a> and
<a class="reference external" href="http://fseoane.net/blog/">Fabian Pedregosa</a>.</li>
<li>New <a class="reference internal" href="modules/mixture.html#dirichlet-process"><em>Dirichlet Process</em></a> Gaussian Mixture
Model by <a class="reference external" href="<http://atpassos.posterous.com>">Alexandre Passos</a></li>
<li><a class="reference internal" href="modules/neighbors.html#neighbors"><em>Nearest Neighbors</em></a> module refactoring by <a class="reference external" href="http://www.astro.washington.edu/users/vanderplas/">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.</li>
<li>Improvements on the <a class="reference internal" href="modules/feature_selection.html#feature-selection"><em>Feature selection</em></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.</li>
<li><a class="reference internal" href="modules/decomposition.html#sparsepca"><em>Sparse Principal Components Analysis (SparsePCA and MiniBatchSparsePCA)</em></a> by <a class="reference external" href="http://vene.ro">Vlad Niculae</a>, <a class="reference external" href="http://gael-varoquaux.info">Gael Varoquaux</a> and
<a class="reference external" href="http://www-sop.inria.fr/members/Alexandre.Gramfort/">Alexandre Gramfort</a></li>
<li>Printing an estimator now behaves independently of architectures
and Python version thanks to Jean Kossaifi.</li>
<li><a class="reference internal" href="datasets/index.html#libsvm-loader"><em>Loader for libsvm/svmlight format</em></a> by
<a class="reference external" href="http://www.mblondel.org/journal/">Mathieu Blondel</a> and <a class="reference external" href="https://github.com/larsmans">Lars Buitinck</a></li>
<li>Documentation improvements: thumbnails in
<a class="reference internal" href="auto_examples/index.html#examples-index"><em>example gallery</em></a> by <a class="reference external" href="http://fseoane.net/blog/">Fabian Pedregosa</a>.</li>
<li>Important bugfixes in <a class="reference internal" href="modules/svm.html#svm"><em>Support Vector Machines</em></a> module (segfaults, bad
performance) by <a class="reference external" href="http://fseoane.net/blog/">Fabian Pedregosa</a>.</li>
<li>Added <a class="reference internal" href="modules/naive_bayes.html#multinomial-naive-bayes"><em>Multinomial Naive Bayes</em></a> and <a class="reference internal" href="modules/naive_bayes.html#bernoulli-naive-bayes"><em>Bernoulli Naive Bayes</em></a>
by <a class="reference external" href="https://github.com/larsmans">Lars Buitinck</a></li>
<li>Text feature extraction optimizations by Lars Buitinck</li>
<li>Chi-Square feature selection
(<tt class="xref py py-func docutils literal"><span class="pre">feature_selection.univariate_selection.chi2</span></tt>) by <cite>Lars Buitinck</cite>.</li>
<li><a class="reference internal" href="datasets/index.html#sample-generators"><em>Sample generators</em></a> module refactoring by <a class="reference external" href="http://www.montefiore.ulg.ac.be/~glouppe/">Gilles Louppe</a></li>
<li><a class="reference internal" href="modules/multiclass.html#multiclass"><em>Multiclass and multilabel algorithms</em></a> by <a class="reference external" href="http://www.mblondel.org/journal/">Mathieu Blondel</a></li>
<li>Ball tree rewrite by <a class="reference external" href="http://www.astro.washington.edu/users/vanderplas/">Jake Vanderplas</a></li>
<li>Implementation of <a class="reference internal" href="modules/clustering.html#dbscan"><em>DBSCAN</em></a> algorithm by Robert Layton</li>
<li>Kmeans predict and transform by Robert Layton</li>
<li>Preprocessing module refactoring by <a class="reference external" href="http://twitter.com/ogrisel">Olivier Grisel</a></li>
<li>Faster mean shift by Conrad Lee</li>
<li>New <a class="reference internal" href="modules/cross_validation.html#bootstrap"><em>Bootstrapping cross-validation</em></a>, <a class="reference internal" href="modules/cross_validation.html#shufflesplit"><em>Random permutations cross-validation a.k.a. Shuffle & Split</em></a> and various other
improvements in cross validation schemes by <a class="reference external" href="http://twitter.com/ogrisel">Olivier Grisel</a> and
<a class="reference external" href="http://gael-varoquaux.info">Gael Varoquaux</a></li>
<li>Adjusted Rand index and V-Measure clustering evaluation metrics by <a class="reference external" href="http://twitter.com/ogrisel">Olivier Grisel</a></li>
<li>Added <a class="reference internal" href="modules/generated/sklearn.linear_model.OrthogonalMatchingPursuit.html#sklearn.linear_model.OrthogonalMatchingPursuit" title="sklearn.linear_model.OrthogonalMatchingPursuit"><tt class="xref py py-class docutils literal"><span class="pre">Orthogonal</span> <span class="pre">Matching</span> <span class="pre">Pursuit</span></tt></a> by <a class="reference external" href="http://vene.ro">Vlad Niculae</a></li>
<li>Added 2D-patch extractor utilites in the <a class="reference internal" href="modules/feature_extraction.html#feature-extraction"><em>Feature extraction</em></a> module by <a class="reference external" href="http://vene.ro">Vlad Niculae</a></li>
<li>Implementation of <a class="reference internal" href="modules/generated/sklearn.linear_model.LassoLarsCV.html#sklearn.linear_model.LassoLarsCV" title="sklearn.linear_model.LassoLarsCV"><tt class="xref py py-class docutils literal"><span class="pre">linear_model.LassoLarsCV</span></tt></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"><tt class="xref py py-class docutils literal"><span class="pre">linear_model.LassoLarsIC</span></tt></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://www-sop.inria.fr/members/Alexandre.Gramfort/">Alexandre Gramfort</a></li>
<li>Scalability improvements to <a class="reference internal" href="modules/generated/sklearn.metrics.roc_curve.html#sklearn.metrics.roc_curve" title="sklearn.metrics.roc_curve"><tt class="xref py py-func docutils literal"><span class="pre">metrics.roc_curve</span></tt></a> by Olivier Hervieu</li>
<li>Distance helper functions <a class="reference internal" href="modules/generated/sklearn.metrics.pairwise.pairwise_distances.html#sklearn.metrics.pairwise.pairwise_distances" title="sklearn.metrics.pairwise.pairwise_distances"><tt class="xref py py-func docutils literal"><span class="pre">metrics.pairwise.pairwise_distances</span></tt></a>
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"><tt class="xref py py-func docutils literal"><span class="pre">metrics.pairwise.pairwise_kernels</span></tt></a> by Robert Layton</li>
<li><a class="reference internal" href="modules/generated/sklearn.cluster.MiniBatchKMeans.html#sklearn.cluster.MiniBatchKMeans" title="sklearn.cluster.MiniBatchKMeans"><tt class="xref py py-class docutils literal"><span class="pre">Mini-Batch</span> <span class="pre">K-Means</span></tt></a> by Nelle Varoquaux and Peter Prettenhofer.</li>
<li><a class="reference internal" href="datasets/index.html#mldata"><em>Downloading datasets from the mldata.org repository</em></a> utilities by Pietro Berkes.</li>
<li><a class="reference internal" href="datasets/index.html#olivetti-faces"><em>The Olivetti faces dataset</em></a> by <a class="reference external" href="http://www-etud.iro.umontreal.ca/~wardefar/">David Warde-Farley</a>.</li>
</ul>
</div></blockquote>
</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 updgrading from scikit-learn
version 0.8:</p>
<blockquote>
<div><ul>
<li><p class="first">The <tt class="docutils literal"><span class="pre">scikits.learn</span></tt> package was renamed <tt class="docutils literal"><span class="pre">sklearn</span></tt>. There is
still a <tt class="docutils literal"><span class="pre">scikits.learn</span></tt> 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-python"><pre>find -name "*.py" | xargs sed -i 's/\bscikits.learn\b/sklearn/g'</pre>
</div>
</li>
<li><p class="first">Estimators no longer accept model parameters as <tt class="docutils literal"><span class="pre">fit</span></tt> arguments:
instead all parameters must be only be passed as constructor
arguments or using the now public <tt class="docutils literal"><span class="pre">set_params</span></tt> method inhereted
from <tt class="xref py py-class docutils literal"><span class="pre">base.BaseEstimator</span></tt>.</p>
<p>Some estimators can still accept keyword arguments on the <tt class="docutils literal"><span class="pre">fit</span></tt>
but this is restricted to data-dependent values (e.g. a Gram matrix
or an affinity matrix that are precomputed from the <tt class="docutils literal"><span class="pre">X</span></tt> data matrix.</p>
</li>
<li><p class="first">The <tt class="docutils literal"><span class="pre">cross_val</span></tt> package has been renamed to <tt class="docutils literal"><span class="pre">cross_validation</span></tt>
although there is also a <tt class="docutils literal"><span class="pre">cross_val</span></tt> 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-python"><pre>find -name "*.py" | xargs sed -i 's/\bcross_val\b/cross_validation/g'</pre>
</div>
</li>
<li><p class="first">The <tt class="docutils literal"><span class="pre">score_func</span></tt> argument of the
<tt class="docutils literal"><span class="pre">sklearn.cross_validation.cross_val_score</span></tt> function is now expected
to accept <tt class="docutils literal"><span class="pre">y_test</span></tt> and <tt class="docutils literal"><span class="pre">y_predicted</span></tt> as only arguments for
classification and regression tasks or <tt class="docutils literal"><span class="pre">X_test</span></tt> for unsupervised
estimators.</p>
</li>
<li><p class="first"><tt class="docutils literal"><span class="pre">gamma</span></tt> parameter for support vector machine algorithms is set
to <tt class="docutils literal"><span class="pre">1</span> <span class="pre">/</span> <span class="pre">n_features</span></tt> by default, instead of <tt class="docutils literal"><span class="pre">1</span> <span class="pre">/</span> <span class="pre">n_samples</span></tt>.</p>
</li>
<li><p class="first">The <tt class="docutils literal"><span class="pre">sklearn.hmm</span></tt> 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 class="first"><tt class="docutils literal"><span class="pre">sklearn.neighbors</span></tt> has been made into a submodule. The two previously
available estimators, <tt class="docutils literal"><span class="pre">NeighborsClassifier</span></tt> and <tt class="docutils literal"><span class="pre">NeighborsRegressor</span></tt>
have been marked as deprecated. Their functionality has been divided
among five new classes: <tt class="docutils literal"><span class="pre">NearestNeighbors</span></tt> for unsupervised neighbors
searches, <tt class="docutils literal"><span class="pre">KNeighborsClassifier</span></tt> & <tt class="docutils literal"><span class="pre">RadiusNeighborsClassifier</span></tt>
for supervised classification problems, and <tt class="docutils literal"><span class="pre">KNeighborsRegressor</span></tt>
& <tt class="docutils literal"><span class="pre">RadiusNeighborsRegressor</span></tt> for supervised regression problems.</p>
</li>
<li><p class="first"><tt class="docutils literal"><span class="pre">sklearn.ball_tree.BallTree</span></tt> has been moved to
<tt class="docutils literal"><span class="pre">sklearn.neighbors.BallTree</span></tt>. Using the former will generate a warning.</p>
</li>
<li><p class="first"><tt class="docutils literal"><span class="pre">sklearn.linear_model.LARS()</span></tt> and related classes (LassoLARS,
LassoLARSCV, etc.) have been renamed to
<tt class="docutils literal"><span class="pre">sklearn.linear_model.Lars()</span></tt>.</p>
</li>
<li><p class="first">All distance metrics and kernels in <tt class="docutils literal"><span class="pre">sklearn.metrics.pairwise</span></tt> 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 class="first"><tt class="docutils literal"><span class="pre">sklearn.metrics.pairwise.l1_distance</span></tt> is now called <tt class="docutils literal"><span class="pre">manhattan_distance</span></tt>,
and by default returns the pairwise distance. For the component wise distance,
set the parameter <tt class="docutils literal"><span class="pre">sum_over_features</span></tt> to <tt class="docutils literal"><span class="pre">False</span></tt>.</p>
</li>
</ul>
</div></blockquote>
<p>Backward compatibilty package aliases and other deprecated classes and
functions will be removed in version 0.11.</p>
</div>
<div class="section" id="id8">
<h2>People<a class="headerlink" href="#id8" title="Permalink to this headline">¶</a></h2>
<p>38 people contributed to this release.</p>
<blockquote>
<div><ul class="simple">
<li>387 <a class="reference external" href="http://vene.ro">Vlad Niculae</a></li>
<li>320 <a class="reference external" href="http://twitter.com/ogrisel">Olivier Grisel</a></li>
<li>192 <a class="reference external" href="https://github.com/larsmans">Lars Buitinck</a></li>
<li>179 <a class="reference external" href="http://gael-varoquaux.info">Gael Varoquaux</a></li>
<li>168 <a class="reference external" href="http://fseoane.net/blog/">Fabian Pedregosa</a> (<a class="reference external" href="http://inria.fr">INRIA</a>, <a class="reference external" href="http://parietal.saclay.inria.fr/">Parietal Team</a>)</li>
<li>127 <a class="reference external" href="http://www.astro.washington.edu/users/vanderplas/">Jake Vanderplas</a></li>
<li>120 <a class="reference external" href="http://www.mblondel.org/journal/">Mathieu Blondel</a></li>
<li>85 <a class="reference external" href="<http://atpassos.posterous.com>">Alexandre Passos</a></li>
<li>67 <a class="reference external" href="http://www-sop.inria.fr/members/Alexandre.Gramfort/">Alexandre Gramfort</a></li>
<li>57 <a class="reference external" href="http://sites.google.com/site/peterprettenhofer/">Peter Prettenhofer</a></li>
<li>56 <a class="reference external" href="http://www.montefiore.ulg.ac.be/~glouppe/">Gilles Louppe</a></li>
<li>42 Robert Layton</li>
<li>38 Nelle Varoquaux</li>
<li>32 Jean Kossaifi</li>
<li>30 Conrad Lee</li>
<li>22 Pietro Berkes</li>
<li>18 andy</li>
<li>17 David Warde-Farley</li>
<li>12 Brian Holt</li>
<li>11 Robert</li>
<li>8 Amit Aides</li>
<li>8 <a class="reference external" href="http://parietal.saclay.inria.fr/Members/virgile-fritsch">Virgile Fritsch</a></li>
<li>7 <a class="reference external" href="http://www.onerussian.com/">Yaroslav Halchenko</a></li>
<li>6 Salvatore Masecchia</li>
<li>5 Paolo Losi</li>
<li>4 Vincent Schut</li>
<li>3 Alexis Metaireau</li>
<li>3 Bryan Silverthorn</li>
<li>3 <a class="reference external" href="http://www.ais.uni-bonn.de/~amueller/">Andreas Müller</a></li>
<li>2 Minwoo Jake Lee</li>
<li>1 Emmanuelle Gouillart</li>
<li>1 Keith Goodman</li>
<li>1 Lucas Wiman</li>
<li>1 <a class="reference external" href="http://pinto.scripts.mit.edu/">Nicolas Pinto</a></li>
<li>1 Thouis (Ray) Jones</li>
<li>1 Tim Sheerman-Chase</li>
</ul>
</div></blockquote>
</div>
</div>
<div class="section" id="changes-0-8">
<span id="id9"></span><h1>0.8<a class="headerlink" href="#changes-0-8" title="Permalink to this headline">¶</a></h1>
<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"><em>Hierarchical clustering</em></a>,
<a class="reference internal" href="modules/pls.html#pls"><em>Partial Least Squares</em></a>, <a class="reference internal" href="modules/decomposition.html#nmf"><em>Non-negative matrix factorization (NMF or NNMF)</em></a>, initial support for Python 3 and by important
enhacements and bug fixes.</p>
<div class="section" id="id10">
<h2>Changelog<a class="headerlink" href="#id10" title="Permalink to this headline">¶</a></h2>
<p>Several new modules where introduced during this release:</p>
<blockquote>
<div><ul class="simple">
<li>New <a class="reference internal" href="modules/clustering.html#hierarchical-clustering"><em>Hierarchical clustering</em></a> module by Vincent Michel,
<a class="reference external" href="http://parietal.saclay.inria.fr/Members/bertrand-thirion">Bertrand Thirion</a>, <a class="reference external" href="http://www-sop.inria.fr/members/Alexandre.Gramfort/">Alexandre Gramfort</a> and <a class="reference external" href="http://gael-varoquaux.info">Gael Varoquaux</a>.</li>
<li><a class="reference internal" href="modules/decomposition.html#kernel-pca"><em>Kernel PCA</em></a> implementation by <a class="reference external" href="http://www.mblondel.org/journal/">Mathieu Blondel</a></li>
<li><a class="reference internal" href="datasets/index.html#labeled-faces-in-the-wild"><em>The Labeled Faces in the Wild face recognition dataset</em></a> by <a class="reference external" href="http://twitter.com/ogrisel">Olivier Grisel</a>.</li>
<li>New <a class="reference internal" href="modules/pls.html#pls"><em>Partial Least Squares</em></a> module by <a class="reference external" href="http://www.lnao.fr/spip.php?rubrique30">Edouard Duchesnay</a>.</li>
<li><a class="reference internal" href="modules/decomposition.html#nmf"><em>Non-negative matrix factorization (NMF or NNMF)</em></a> module <a class="reference external" href="http://vene.ro">Vlad Niculae</a></li>
<li>Implementation of the <a class="reference internal" href="modules/covariance.html#oracle-approximating-shrinkage"><em>Oracle Approximating Shrinkage</em></a> algorithm by
<a class="reference external" href="http://parietal.saclay.inria.fr/Members/virgile-fritsch">Virgile Fritsch</a> in the <a class="reference internal" href="modules/covariance.html#covariance"><em>Covariance estimation</em></a> module.</li>
</ul>
</div></blockquote>
<p>Some other modules benefited from significant improvements or cleanups.</p>
<blockquote>
<div><ul class="simple">
<li>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://fseoane.net/blog/">Fabian Pedregosa</a>.</li>
<li><a class="reference internal" href="modules/generated/sklearn.decomposition.PCA.html#sklearn.decomposition.PCA" title="sklearn.decomposition.PCA"><tt class="xref py py-class docutils literal"><span class="pre">decomposition.PCA</span></tt></a> is now usable from the Pipeline object by <a class="reference external" href="http://twitter.com/ogrisel">Olivier Grisel</a>.</li>
<li>Guide <a class="reference internal" href="developers/performance.html#performance-howto"><em>How to optimize for speed</em></a> by <a class="reference external" href="http://twitter.com/ogrisel">Olivier Grisel</a>.</li>
<li>Fixes for memory leaks in libsvm bindings, 64-bit safer BallTree by Lars Buitinck.</li>
<li>bug and style fixing in <a class="reference internal" href="modules/clustering.html#k-means"><em>K-means</em></a> algorithm by Jan Schlüter.</li>
<li>Add attribute coverged to Gaussian Mixture Models by Vincent Schut.</li>
<li>Implement <cite>transform</cite>, <cite>predict_log_proba</cite> in <a class="reference internal" href="modules/generated/sklearn.lda.LDA.html#sklearn.lda.LDA" title="sklearn.lda.LDA"><tt class="xref py py-class docutils literal"><span class="pre">lda.LDA</span></tt></a> by <a class="reference external" href="http://www.mblondel.org/journal/">Mathieu Blondel</a>.</li>
<li>Refactoring in the <a class="reference internal" href="modules/svm.html#svm"><em>Support Vector Machines</em></a> module and bug fixes by <a class="reference external" href="http://fseoane.net/blog/">Fabian Pedregosa</a>,
<a class="reference external" href="http://gael-varoquaux.info">Gael Varoquaux</a> and Amit Aides.</li>
<li>Refactored SGD module (removed code duplication, better variable naming),
added interface for sample weight by <a class="reference external" href="http://sites.google.com/site/peterprettenhofer/">Peter Prettenhofer</a>.</li>
<li>Wrapped BallTree with Cython by Thouis (Ray) Jones.</li>
<li>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"><tt class="xref py py-func docutils literal"><span class="pre">svm.l1_min_c</span></tt></a> by Paolo Losi.</li>
<li>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="http://twitter.com/ogrisel">Olivier Grisel</a>, Yann Malet, <a class="reference external" href="http://pinto.scripts.mit.edu/">Nicolas Pinto</a>, Lars Buitinck and
<a class="reference external" href="http://fseoane.net/blog/">Fabian Pedregosa</a>.</li>
</ul>
</div></blockquote>
</div>
<div class="section" id="id11">
<h2>People<a class="headerlink" href="#id11" title="Permalink to this headline">¶</a></h2>
<p>People that made this release possible preceeded by number of commits:</p>
<blockquote>
<div><ul class="simple">
<li>159 <a class="reference external" href="http://twitter.com/ogrisel">Olivier Grisel</a></li>
<li>96 <a class="reference external" href="http://gael-varoquaux.info">Gael Varoquaux</a></li>
<li>96 <a class="reference external" href="http://vene.ro">Vlad Niculae</a></li>
<li>94 <a class="reference external" href="http://fseoane.net/blog/">Fabian Pedregosa</a></li>
<li>36 <a class="reference external" href="http://www-sop.inria.fr/members/Alexandre.Gramfort/">Alexandre Gramfort</a></li>
<li>32 Paolo Losi</li>
<li>31 <a class="reference external" href="http://www.lnao.fr/spip.php?rubrique30">Edouard Duchesnay</a></li>
<li>30 <a class="reference external" href="http://www.mblondel.org/journal/">Mathieu Blondel</a></li>
<li>25 <a class="reference external" href="http://sites.google.com/site/peterprettenhofer/">Peter Prettenhofer</a></li>
<li>22 <a class="reference external" href="http://pinto.scripts.mit.edu/">Nicolas Pinto</a></li>
<li>11 <a class="reference external" href="http://parietal.saclay.inria.fr/Members/virgile-fritsch">Virgile Fritsch</a></li>
<li>7 Lars Buitinck</li>
<li>6 Vincent Michel</li>
<li>5 <a class="reference external" href="http://parietal.saclay.inria.fr/Members/bertrand-thirion">Bertrand Thirion</a></li>
<li>4 Thouis (Ray) Jones</li>
<li>4 Vincent Schut</li>
<li>3 Jan Schlüter</li>
<li>2 Julien Miotte</li>
<li>2 <a class="reference external" href="http://www.lnao.fr/spip.php?rubrique19">Matthieu Perrot</a></li>
<li>2 Yann Malet</li>
<li>2 <a class="reference external" href="http://www.onerussian.com/">Yaroslav Halchenko</a></li>
<li>1 Amit Aides</li>
<li>1 <a class="reference external" href="http://www.ais.uni-bonn.de/~amueller/">Andreas Müller</a></li>
<li>1 Feth Arezki</li>
<li>1 Meng Xinfan</li>
</ul>
</div></blockquote>
</div>
</div>
<div class="section" id="changes-0-7">
<span id="id12"></span><h1>0.7<a class="headerlink" href="#changes-0-7" title="Permalink to this headline">¶</a></h1>
<p>scikit-learn 0.7 was released in March 2011, roughly three months
after the 0.6 release. This release is marked by the speed
improvements in existing algorithms like k-Nearest Neighbors and
K-Means algorithm and by the inclusion of an efficient algorithm for
computing the Ridge Generalized Cross Validation solution. Unlike the
preceding release, no new modules where added to this release.</p>
<div class="section" id="id13">
<h2>Changelog<a class="headerlink" href="#id13" title="Permalink to this headline">¶</a></h2>
<blockquote>
<div><ul class="simple">
<li>Performance improvements for Gaussian Mixture Model sampling [Jan
Schlüter].</li>
<li>Implementation of efficient leave-one-out cross-validated Ridge in
<a class="reference internal" href="modules/generated/sklearn.linear_model.RidgeCV.html#sklearn.linear_model.RidgeCV" title="sklearn.linear_model.RidgeCV"><tt class="xref py py-class docutils literal"><span class="pre">linear_model.RidgeCV</span></tt></a> [<a class="reference external" href="http://www.mblondel.org/journal/">Mathieu Blondel</a>]</li>
<li>Better handling of collinearity and early stopping in
<a class="reference internal" href="modules/generated/sklearn.linear_model.lars_path.html#sklearn.linear_model.lars_path" title="sklearn.linear_model.lars_path"><tt class="xref py py-func docutils literal"><span class="pre">linear_model.lars_path</span></tt></a> [<a class="reference external" href="http://www-sop.inria.fr/members/Alexandre.Gramfort/">Alexandre Gramfort</a> and <a class="reference external" href="http://fseoane.net/blog/">Fabian
Pedregosa</a>].</li>
<li>Fixes for liblinear ordering of labels and sign of coefficients
[Dan Yamins, Paolo Losi, <a class="reference external" href="http://www.mblondel.org/journal/">Mathieu Blondel</a> and <a class="reference external" href="http://fseoane.net/blog/">Fabian Pedregosa</a>].</li>
<li>Performance improvements for Nearest Neighbors algorithm in
high-dimensional spaces [<a class="reference external" href="http://fseoane.net/blog/">Fabian Pedregosa</a>].</li>
<li>Performance improvements for <a class="reference internal" href="modules/generated/sklearn.cluster.KMeans.html#sklearn.cluster.KMeans" title="sklearn.cluster.KMeans"><tt class="xref py py-class docutils literal"><span class="pre">cluster.KMeans</span></tt></a> [<a class="reference external" href="http://gael-varoquaux.info">Gael
Varoquaux</a> and <a class="reference external" href="http://www-etud.iro.umontreal.ca/~bergstrj/">James Bergstra</a>].</li>
<li>Sanity checks for SVM-based classes [<a class="reference external" href="http://www.mblondel.org/journal/">Mathieu Blondel</a>].</li>
<li>Refactoring of <tt class="xref py py-class docutils literal"><span class="pre">neighbors.NeighborsClassifier</span></tt> and
<a class="reference internal" href="modules/generated/sklearn.neighbors.kneighbors_graph.html#sklearn.neighbors.kneighbors_graph" title="sklearn.neighbors.kneighbors_graph"><tt class="xref py py-func docutils literal"><span class="pre">neighbors.kneighbors_graph</span></tt></a>: added different algorithms for
the k-Nearest Neighbor Search and implemented a more stable
algorithm for finding barycenter weigths. Also added some
developer documentation for this module, see
<a class="reference external" href="https://github.com/scikit-learn/scikit-learn/wiki/Neighbors-working-notes">notes_neighbors</a> for more information [<a class="reference external" href="http://fseoane.net/blog/">Fabian Pedregosa</a>].</li>
<li>Documentation improvements: Added <tt class="xref py py-class docutils literal"><span class="pre">pca.RandomizedPCA</span></tt> and
<a class="reference internal" href="modules/generated/sklearn.linear_model.LogisticRegression.html#sklearn.linear_model.LogisticRegression" title="sklearn.linear_model.LogisticRegression"><tt class="xref py py-class docutils literal"><span class="pre">linear_model.LogisticRegression</span></tt></a> to the class
reference. Also added references of matrices used for clustering
and other fixes [<a class="reference external" href="http://gael-varoquaux.info">Gael Varoquaux</a>, <a class="reference external" href="http://fseoane.net/blog/">Fabian Pedregosa</a>, <a class="reference external" href="http://www.mblondel.org/journal/">Mathieu
Blondel</a>, <a class="reference external" href="http://twitter.com/ogrisel">Olivier Grisel</a>, Virgile Fritsch , Emmanuelle
Gouillart]</li>
<li>Binded decision_function in classes that make use of <a class="reference external" href="http://www.csie.ntu.edu.tw/~cjlin/liblinear/">liblinear</a>,
dense and sparse variants, like <a class="reference internal" href="modules/generated/sklearn.svm.LinearSVC.html#sklearn.svm.LinearSVC" title="sklearn.svm.LinearSVC"><tt class="xref py py-class docutils literal"><span class="pre">svm.LinearSVC</span></tt></a> or
<a class="reference internal" href="modules/generated/sklearn.linear_model.LogisticRegression.html#sklearn.linear_model.LogisticRegression" title="sklearn.linear_model.LogisticRegression"><tt class="xref py py-class docutils literal"><span class="pre">linear_model.LogisticRegression</span></tt></a> [<a class="reference external" href="http://fseoane.net/blog/">Fabian Pedregosa</a>].</li>
<li>Performance and API improvements to
<tt class="xref py py-func docutils literal"><span class="pre">metrics.euclidean_distances</span></tt> and to
<tt class="xref py py-class docutils literal"><span class="pre">pca.RandomizedPCA</span></tt> [<a class="reference external" href="http://www-etud.iro.umontreal.ca/~bergstrj/">James Bergstra</a>].</li>
<li>Fix compilation issues under NetBSD [Kamel Ibn Hassen Derouiche]</li>
<li>Allow input sequences of different lengths in <a class="reference internal" href="modules/generated/sklearn.hmm.GaussianHMM.html#sklearn.hmm.GaussianHMM" title="sklearn.hmm.GaussianHMM"><tt class="xref py py-class docutils literal"><span class="pre">hmm.GaussianHMM</span></tt></a>
[<a class="reference external" href="http://www.ee.columbia.edu/~ronw/">Ron Weiss</a>].</li>
<li>Fix bug in affinity propagation caused by incorrect indexing [Xinfan Meng]</li>
</ul>
</div></blockquote>
</div>
<div class="section" id="id14">
<h2>People<a class="headerlink" href="#id14" title="Permalink to this headline">¶</a></h2>
<p>People that made this release possible preceeded by number of commits:</p>
<blockquote>
<div><ul class="simple">
<li>85 <a class="reference external" href="http://fseoane.net/blog/">Fabian Pedregosa</a></li>
<li>67 <a class="reference external" href="http://www.mblondel.org/journal/">Mathieu Blondel</a></li>
<li>20 <a class="reference external" href="http://www-sop.inria.fr/members/Alexandre.Gramfort/">Alexandre Gramfort</a></li>
<li>19 <a class="reference external" href="http://www-etud.iro.umontreal.ca/~bergstrj/">James Bergstra</a></li>
<li>14 Dan Yamins</li>
<li>13 <a class="reference external" href="http://twitter.com/ogrisel">Olivier Grisel</a></li>
<li>12 <a class="reference external" href="http://gael-varoquaux.info">Gael Varoquaux</a></li>
<li>4 Edouard Duchesnay</li>
<li>4 <a class="reference external" href="http://www.ee.columbia.edu/~ronw/">Ron Weiss</a></li>
<li>2 Satrajit Ghosh</li>
<li>2 Vincent Dubourg</li>
<li>1 Emmanuelle Gouillart</li>
<li>1 Kamel Ibn Hassen Derouiche</li>
<li>1 Paolo Losi</li>
<li>1 VirgileFritsch</li>
<li>1 <a class="reference external" href="http://www.onerussian.com/">Yaroslav Halchenko</a></li>
<li>1 Xinfan Meng</li>
</ul>
</div></blockquote>
</div>
</div>
<div class="section" id="changes-0-6">
<span id="id15"></span><h1>0.6<a class="headerlink" href="#changes-0-6" title="Permalink to this headline">¶</a></h1>
<p>scikit-learn 0.6 was released on december 2010. It is marked by the
inclusion of several new modules and a general renaming of old
ones. It is also marked by the inclusion of new example, including
applications to real-world datasets.</p>
<div class="section" id="id16">
<h2>Changelog<a class="headerlink" href="#id16" title="Permalink to this headline">¶</a></h2>
<blockquote>
<div><ul class="simple">
<li>New <a class="reference external" href="http://scikit-learn.org/stable/modules/sgd.html">stochastic gradient</a> descent
module by Peter Prettenhofer. The module comes with complete
documentation and examples.</li>
<li>Improved svm module: memory consumption has been reduced by 50%,
heuristic to automatically set class weights, possibility to
assign weights to samples (see
<a class="reference internal" href="auto_examples/svm/plot_weighted_samples.html#example-svm-plot-weighted-samples-py"><em>SVM: Weighted samples</em></a> for an example).</li>
<li>New <a class="reference internal" href="modules/gaussian_process.html#gaussian-process"><em>Gaussian Processes</em></a> module by Vincent Dubourg. This module
also has great documentation and some very neat examples. See
<a class="reference internal" href="auto_examples/gaussian_process/plot_gp_regression.html#example-gaussian-process-plot-gp-regression-py"><em>Gaussian Processes regression: basic introductory example</em></a> or
<a class="reference internal" href="auto_examples/gaussian_process/plot_gp_probabilistic_classification_after_regression.html#example-gaussian-process-plot-gp-probabilistic-classification-after-regression-py"><em>Gaussian Processes classification example: exploiting the probabilistic output</em></a>
for a taste of what can be done.</li>
<li>It is now possible to use liblinear’s Multi-class SVC (option
multi_class in <a class="reference internal" href="modules/generated/sklearn.svm.LinearSVC.html#sklearn.svm.LinearSVC" title="sklearn.svm.LinearSVC"><tt class="xref py py-class docutils literal"><span class="pre">svm.LinearSVC</span></tt></a>)</li>
<li>New features and performance improvements of text feature
extraction.</li>
<li>Improved sparse matrix support, both in main classes
(<a class="reference internal" href="modules/generated/sklearn.grid_search.GridSearchCV.html#sklearn.grid_search.GridSearchCV" title="sklearn.grid_search.GridSearchCV"><tt class="xref py py-class docutils literal"><span class="pre">grid_search.GridSearchCV</span></tt></a>) as in modules
sklearn.svm.sparse and sklearn.linear_model.sparse.</li>
<li>Lots of cool new examples and a new section that uses real-world
datasets was created. These include:
<a class="reference internal" href="auto_examples/applications/plot_face_recognition.html#example-applications-plot-face-recognition-py"><em>Faces recognition example using eigenfaces and SVMs</em></a>,
<a class="reference internal" href="auto_examples/applications/plot_species_distribution_modeling.html#example-applications-plot-species-distribution-modeling-py"><em>Species distribution modeling</em></a>,
<a class="reference internal" href="auto_examples/applications/svm_gui.html#example-applications-svm-gui-py"><em>Libsvm GUI</em></a>,
<a class="reference internal" href="auto_examples/applications/wikipedia_principal_eigenvector.html#example-applications-wikipedia-principal-eigenvector-py"><em>Wikipedia princial eigenvector</em></a> and
others.</li>
<li>Faster <a class="reference internal" href="modules/linear_model.html#least-angle-regression"><em>Least Angle Regression</em></a> algorithm. It is now 2x
faster than the R version on worst case and up to 10x times faster
on some cases.</li>
<li>Faster coordinate descent algorithm. In particular, the full path
version of lasso (<a class="reference internal" href="modules/generated/sklearn.linear_model.lasso_path.html#sklearn.linear_model.lasso_path" title="sklearn.linear_model.lasso_path"><tt class="xref py py-func docutils literal"><span class="pre">linear_model.lasso_path</span></tt></a>) is more than
200x times faster than before.</li>
<li>It is now possible to get probability estimates from a
<a class="reference internal" href="modules/generated/sklearn.linear_model.LogisticRegression.html#sklearn.linear_model.LogisticRegression" title="sklearn.linear_model.LogisticRegression"><tt class="xref py py-class docutils literal"><span class="pre">linear_model.LogisticRegression</span></tt></a> model.</li>
<li>module renaming: the glm module has been renamed to linear_model,
the gmm module has been included into the more general mixture
model and the sgd module has been included in linear_model.</li>
<li>Lots of bug fixes and documentation improvements.</li>
</ul>
</div></blockquote>
</div>
<div class="section" id="id17">
<h2>People<a class="headerlink" href="#id17" title="Permalink to this headline">¶</a></h2>
<p>People that made this release possible preceeded by number of commits:</p>
<blockquote>
<div><ul class="simple">
<li>207 <a class="reference external" href="http://twitter.com/ogrisel">Olivier Grisel</a></li>
<li>167 <a class="reference external" href="http://fseoane.net/blog/">Fabian Pedregosa</a></li>
<li>97 <a class="reference external" href="http://sites.google.com/site/peterprettenhofer/">Peter Prettenhofer</a></li>
<li>68 <a class="reference external" href="http://www-sop.inria.fr/members/Alexandre.Gramfort/">Alexandre Gramfort</a></li>
<li>59 <a class="reference external" href="http://www.mblondel.org/journal/">Mathieu Blondel</a></li>
<li>55 <a class="reference external" href="http://gael-varoquaux.info">Gael Varoquaux</a></li>
<li>33 Vincent Dubourg</li>
<li>21 <a class="reference external" href="http://www.ee.columbia.edu/~ronw/">Ron Weiss</a></li>
<li>9 Bertrand Thirion</li>
<li>3 <a class="reference external" href="<http://atpassos.posterous.com>">Alexandre Passos</a></li>
<li>3 Anne-Laure Fouque</li>
<li>2 Ronan Amicel</li>
<li>1 <a class="reference external" href="http://osdf.github.com/">Christian Osendorfer</a></li>
</ul>
</div></blockquote>
</div>
</div>
<div class="section" id="changes-0-5">
<span id="id18"></span><h1>0.5<a class="headerlink" href="#changes-0-5" title="Permalink to this headline">¶</a></h1>
<div class="section" id="id19">
<h2>Changelog<a class="headerlink" href="#id19" title="Permalink to this headline">¶</a></h2>
<div class="section" id="new-classes">
<h3>New classes<a class="headerlink" href="#new-classes" title="Permalink to this headline">¶</a></h3>
<blockquote>
<div><ul class="simple">
<li>Support for sparse matrices in some classifiers of modules
<tt class="docutils literal"><span class="pre">svm</span></tt> and <tt class="docutils literal"><span class="pre">linear_model</span></tt> (see <tt class="xref py py-class docutils literal"><span class="pre">svm.sparse.SVC</span></tt>,
<tt class="xref py py-class docutils literal"><span class="pre">svm.sparse.SVR</span></tt>, <tt class="xref py py-class docutils literal"><span class="pre">svm.sparse.LinearSVC</span></tt>,
<a class="reference internal" href="modules/generated/sklearn.linear_model.sparse.Lasso.html#sklearn.linear_model.sparse.Lasso" title="sklearn.linear_model.sparse.Lasso"><tt class="xref py py-class docutils literal"><span class="pre">linear_model.sparse.Lasso</span></tt></a>, <a class="reference internal" href="modules/generated/sklearn.linear_model.sparse.ElasticNet.html#sklearn.linear_model.sparse.ElasticNet" title="sklearn.linear_model.sparse.ElasticNet"><tt class="xref py py-class docutils literal"><span class="pre">linear_model.sparse.ElasticNet</span></tt></a>)</li>
<li>New <a class="reference internal" href="modules/generated/sklearn.pipeline.Pipeline.html#sklearn.pipeline.Pipeline" title="sklearn.pipeline.Pipeline"><tt class="xref py py-class docutils literal"><span class="pre">pipeline.Pipeline</span></tt></a> object to compose different estimators.</li>
<li>Recursive Feature Elimination routines in module
<a class="reference internal" href="modules/feature_selection.html#feature-selection"><em>Feature selection</em></a>.</li>
<li>Addition of various classes capable of cross validation in the
linear_model module (<a class="reference internal" href="modules/generated/sklearn.linear_model.LassoCV.html#sklearn.linear_model.LassoCV" title="sklearn.linear_model.LassoCV"><tt class="xref py py-class docutils literal"><span class="pre">linear_model.LassoCV</span></tt></a>, <a class="reference internal" href="modules/generated/sklearn.linear_model.ElasticNetCV.html#sklearn.linear_model.ElasticNetCV" title="sklearn.linear_model.ElasticNetCV"><tt class="xref py py-class docutils literal"><span class="pre">linear_model.ElasticNetCV</span></tt></a>,
etc.).</li>
<li>New, more efficient LARS algorithm implementation. The Lasso
variant of the algorithm is also implemented. See
<a class="reference internal" href="modules/generated/sklearn.linear_model.lars_path.html#sklearn.linear_model.lars_path" title="sklearn.linear_model.lars_path"><tt class="xref py py-class docutils literal"><span class="pre">linear_model.lars_path</span></tt></a>, <a class="reference internal" href="modules/generated/sklearn.linear_model.Lars.html#sklearn.linear_model.Lars" title="sklearn.linear_model.Lars"><tt class="xref py py-class docutils literal"><span class="pre">linear_model.Lars</span></tt></a> and
<a class="reference internal" href="modules/generated/sklearn.linear_model.LassoLars.html#sklearn.linear_model.LassoLars" title="sklearn.linear_model.LassoLars"><tt class="xref py py-class docutils literal"><span class="pre">linear_model.LassoLars</span></tt></a>.</li>
<li>New Hidden Markov Models module (see classes
<a class="reference internal" href="modules/generated/sklearn.hmm.GaussianHMM.html#sklearn.hmm.GaussianHMM" title="sklearn.hmm.GaussianHMM"><tt class="xref py py-class docutils literal"><span class="pre">hmm.GaussianHMM</span></tt></a>, <a class="reference internal" href="modules/generated/sklearn.hmm.MultinomialHMM.html#sklearn.hmm.MultinomialHMM" title="sklearn.hmm.MultinomialHMM"><tt class="xref py py-class docutils literal"><span class="pre">hmm.MultinomialHMM</span></tt></a>,
<a class="reference internal" href="modules/generated/sklearn.hmm.GMMHMM.html#sklearn.hmm.GMMHMM" title="sklearn.hmm.GMMHMM"><tt class="xref py py-class docutils literal"><span class="pre">hmm.GMMHMM</span></tt></a>)</li>
<li>New module feature_extraction (see <a class="reference internal" href="modules/classes.html#feature-extraction-ref"><em>class reference</em></a>)</li>
<li>New FastICA algorithm in module sklearn.fastica</li>
</ul>
</div></blockquote>
</div>
<div class="section" id="documentation">
<h3>Documentation<a class="headerlink" href="#documentation" title="Permalink to this headline">¶</a></h3>
<blockquote>
<div><ul class="simple">
<li>Improved documentation for many modules, now separating
narrative documentation from the class reference. As an example,
see <a class="reference external" href="http://scikit-learn.org/stable/modules/svm.html">documentation for the SVM module</a> and the
complete <a class="reference external" href="http://scikit-learn.org/stable/modules/classes.html">class reference</a>.</li>
</ul>
</div></blockquote>
</div>
<div class="section" id="fixes">
<h3>Fixes<a class="headerlink" href="#fixes" title="Permalink to this headline">¶</a></h3>
<blockquote>
<div><ul class="simple">
<li>API changes: adhere variable names to PEP-8, give more
meaningful names.</li>
<li>Fixes for svm module to run on a shared memory context
(multiprocessing).</li>
<li>It is again possible to generate latex (and thus PDF) from the
sphinx docs.</li>
</ul>
</div></blockquote>
</div>
<div class="section" id="examples">
<h3>Examples<a class="headerlink" href="#examples" title="Permalink to this headline">¶</a></h3>
<blockquote>
<div><ul class="simple">
<li>new examples using some of the mlcomp datasets:
<a class="reference internal" href="auto_examples/mlcomp_sparse_document_classification.html#example-mlcomp-sparse-document-classification-py"><em>Classification of text documents: using a MLComp dataset</em></a>,
<a class="reference internal" href="auto_examples/document_classification_20newsgroups.html#example-document-classification-20newsgroups-py"><em>Classification of text documents using sparse features</em></a></li>
<li>Many more examaples. <a class="reference external" href="http://scikit-learn.org/stable/auto_examples/index.html">See here</a>
the full list of examples.</li>
</ul>
</div></blockquote>
</div>
<div class="section" id="external-dependencies">
<h3>External dependencies<a class="headerlink" href="#external-dependencies" title="Permalink to this headline">¶</a></h3>
<blockquote>
<div><ul class="simple">
<li>Joblib is now a dependencie of this package, although it is
shipped with (sklearn.externals.joblib).</li>
</ul>
</div></blockquote>
</div>
<div class="section" id="removed-modules">
<h3>Removed modules<a class="headerlink" href="#removed-modules" title="Permalink to this headline">¶</a></h3>
<blockquote>
<div><ul class="simple">
<li>Module ann (Artificial Neural Networks) has been removed from
the distribution. Users wanting this sort of algorithms should
take a look into pybrain.</li>
</ul>
</div></blockquote>
</div>
<div class="section" id="misc">
<h3>Misc<a class="headerlink" href="#misc" title="Permalink to this headline">¶</a></h3>
<blockquote>
<div><ul class="simple">
<li>New sphinx theme for the web page.</li>
</ul>
</div></blockquote>
</div>
</div>
<div class="section" id="authors">
<h2>Authors<a class="headerlink" href="#authors" title="Permalink to this headline">¶</a></h2>
<p>The following is a list of authors for this release, preceeded by
number of commits:</p>
<blockquote>
<div><ul class="simple">
<li>262 Fabian Pedregosa</li>
<li>240 Gael Varoquaux</li>
<li>149 Alexandre Gramfort</li>
<li>116 Olivier Grisel</li>
<li>40 Vincent Michel</li>
<li>38 Ron Weiss</li>
<li>23 Matthieu Perrot</li>
<li>10 Bertrand Thirion</li>
<li>7 Yaroslav Halchenko</li>
<li>9 VirgileFritsch</li>
<li>6 Edouard Duchesnay</li>
<li>4 Mathieu Blondel</li>
<li>1 Ariel Rokem</li>
<li>1 Matthieu Brucher</li>
</ul>
</div></blockquote>
</div>
</div>
<div class="section" id="id20">
<h1>0.4<a class="headerlink" href="#id20" title="Permalink to this headline">¶</a></h1>
<div class="section" id="id21">
<h2>Changelog<a class="headerlink" href="#id21" title="Permalink to this headline">¶</a></h2>
<p>Major changes in this release include:</p>
<blockquote>
<div><ul class="simple">
<li>Coordinate Descent algorithm (Lasso, ElasticNet) refactoring &
speed improvements (roughly 100x times faster).</li>
<li>Coordinate Descent Refactoring (and bug fixing) for consistency
with R’s package GLMNET.</li>
<li>New metrics module.</li>
<li>New GMM module contributed by Ron Weiss.</li>
<li>Implementation of the LARS algorithm (without Lasso variant for now).</li>
<li>feature_selection module redesign.</li>
<li>Migration to GIT as content management system.</li>
<li>Removal of obsolete attrselect module.</li>
<li>Rename of private compiled extensions (aded underscore).</li>
<li>Removal of legacy unmaintained code.</li>
<li>Documentation improvements (both docstring and rst).</li>
<li>Improvement of the build system to (optionally) link with MKL.
Also, provide a lite BLAS implementation in case no system-wide BLAS is
found.</li>
<li>Lots of new examples.</li>
<li>Many, many bug fixes ...</li>
</ul>
</div></blockquote>
</div>
<div class="section" id="id22">
<h2>Authors<a class="headerlink" href="#id22" title="Permalink to this headline">¶</a></h2>
<p>The committer list for this release is the following (preceded by number
of commits):</p>
<blockquote>
<div><ul class="simple">
<li>143 Fabian Pedregosa</li>
<li>35 Alexandre Gramfort</li>
<li>34 Olivier Grisel</li>
<li>11 Gael Varoquaux</li>
<li>5 Yaroslav Halchenko</li>
<li>2 Vincent Michel</li>
<li>1 Chris Filo Gorgolewski</li>
</ul>
</div></blockquote>
</div>
</div>
</div>
</div>
</div>
<div class="clearer"></div>
</div>
</div>
<div class="footer">
© 2010–2011, scikit-learn developers (BSD License).
Created using <a href="http://sphinx.pocoo.org/">Sphinx</a> 1.1.2. Design by <a href="http://webylimonada.com">Web y Limonada</a>.
<span style="padding-left: 5ex;">
<a href="_sources/whats_new.txt"
rel="nofollow">Show this page source</a>
</span>
</div>