-
Notifications
You must be signed in to change notification settings - Fork 47
/
maxent.html
1708 lines (1384 loc) · 199 KB
/
maxent.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>
<head>
<meta charset="utf-8">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<meta name="theme-color" content="#2D2D2D" />
<title>NLTK :: nltk.classify.maxent</title>
<link rel="stylesheet" href="../../../_static/css/nltk_theme.css"/>
<link rel="stylesheet" href="../../../_static/css/custom.css"/>
<script type="text/javascript" id="documentation_options" data-url_root="../../../" src="../../../_static/documentation_options.js"></script>
<script type="text/javascript" src="../../../_static/documentation_options.js"></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>
<script src="https://email.tl.fortawesome.com/c/eJxNjUEOgyAQAF8jR7Kw6wIHDh7sP1Cw2mgxgmn6-3JsMqc5zEQfE8dkxOY1KKMUOI3ACFKRJpSW2AAp7ontYIaxI6i7XPJVwyeVfCQ550Os3jLrGSNOLgbdAy6s0PBk2TFNjEbsfq31LB0OnX407pJa5v2faRadwSW63mn5KuLyR9j2tgx3zecanl-55R_-jjPs"></script>
</head>
<body>
<div id="nltk-theme-container">
<header>
<div id="logo-container">
<h1>
<a href="../../../index.html">NLTK</a>
</h1>
</div>
<div id="project-container">
<h1>Documentation</h1>
</div>
<a id="menu-toggle" class="fa fa-bars" aria-hidden="true"></a>
<script type="text/javascript">
$("#menu-toggle").click(function() {
$("#menu-toggle").toggleClass("toggled");
$("#side-menu-container").slideToggle(300);
});
</script>
</header>
<div id="content-container">
<div id="side-menu-container">
<div id="search" role="search">
<form id="rtd-search-form" class="wy-form" action="../../../search.html" method="get">
<input type="text" name="q" placeholder="Search" />
<input type="hidden" name="check_keywords" value="yes" />
<input type="hidden" name="area" value="default" />
</form>
</div>
<div id="side-menu" role="navigation">
<p class="caption" role="heading"><span class="caption-text">NLTK Documentation</span></p>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../../../api/nltk.html">API Reference</a></li>
<li class="toctree-l1"><a class="reference external" href="http://www.nltk.org/howto">Example Usage</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../../py-modindex.html">Module Index</a></li>
<li class="toctree-l1"><a class="reference external" href="https://github.com/nltk/nltk/wiki">Wiki</a></li>
<li class="toctree-l1"><a class="reference external" href="https://github.com/nltk/nltk/wiki/FAQ">FAQ</a></li>
</ul>
<p class="caption" role="heading"><span class="caption-text">Installation</span></p>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../../../install.html">Installing NLTK</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../../data.html">Installing NLTK Data</a></li>
</ul>
<p class="caption" role="heading"><span class="caption-text">More</span></p>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../../../news.html">Release Notes</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../../contribute.html">Contributing to NLTK</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../../team.html">NLTK Team</a></li>
</ul>
</div>
</div>
<div id="main-content-container">
<div id="main-content" role="main">
<h1>Source code for nltk.classify.maxent</h1><div class="highlight"><pre>
<span></span><span class="c1"># Natural Language Toolkit: Maximum Entropy Classifiers</span>
<span class="c1">#</span>
<span class="c1"># Copyright (C) 2001-2021 NLTK Project</span>
<span class="c1"># Author: Edward Loper <edloper@gmail.com></span>
<span class="c1"># Dmitry Chichkov <dchichkov@gmail.com> (TypedMaxentFeatureEncoding)</span>
<span class="c1"># URL: <http://nltk.org/></span>
<span class="c1"># For license information, see LICENSE.TXT</span>
<span class="sd">"""</span>
<span class="sd">A classifier model based on maximum entropy modeling framework. This</span>
<span class="sd">framework considers all of the probability distributions that are</span>
<span class="sd">empirically consistent with the training data; and chooses the</span>
<span class="sd">distribution with the highest entropy. A probability distribution is</span>
<span class="sd">"empirically consistent" with a set of training data if its estimated</span>
<span class="sd">frequency with which a class and a feature vector value co-occur is</span>
<span class="sd">equal to the actual frequency in the data.</span>
<span class="sd">Terminology: 'feature'</span>
<span class="sd">======================</span>
<span class="sd">The term *feature* is usually used to refer to some property of an</span>
<span class="sd">unlabeled token. For example, when performing word sense</span>
<span class="sd">disambiguation, we might define a ``'prevword'`` feature whose value is</span>
<span class="sd">the word preceding the target word. However, in the context of</span>
<span class="sd">maxent modeling, the term *feature* is typically used to refer to a</span>
<span class="sd">property of a "labeled" token. In order to prevent confusion, we</span>
<span class="sd">will introduce two distinct terms to disambiguate these two different</span>
<span class="sd">concepts:</span>
<span class="sd"> - An "input-feature" is a property of an unlabeled token.</span>
<span class="sd"> - A "joint-feature" is a property of a labeled token.</span>
<span class="sd">In the rest of the ``nltk.classify`` module, the term "features" is</span>
<span class="sd">used to refer to what we will call "input-features" in this module.</span>
<span class="sd">In literature that describes and discusses maximum entropy models,</span>
<span class="sd">input-features are typically called "contexts", and joint-features</span>
<span class="sd">are simply referred to as "features".</span>
<span class="sd">Converting Input-Features to Joint-Features</span>
<span class="sd">-------------------------------------------</span>
<span class="sd">In maximum entropy models, joint-features are required to have numeric</span>
<span class="sd">values. Typically, each input-feature ``input_feat`` is mapped to a</span>
<span class="sd">set of joint-features of the form:</span>
<span class="sd">| joint_feat(token, label) = { 1 if input_feat(token) == feat_val</span>
<span class="sd">| { and label == some_label</span>
<span class="sd">| {</span>
<span class="sd">| { 0 otherwise</span>
<span class="sd">For all values of ``feat_val`` and ``some_label``. This mapping is</span>
<span class="sd">performed by classes that implement the ``MaxentFeatureEncodingI``</span>
<span class="sd">interface.</span>
<span class="sd">"""</span>
<span class="k">try</span><span class="p">:</span>
<span class="kn">import</span> <span class="nn">numpy</span>
<span class="k">except</span> <span class="ne">ImportError</span><span class="p">:</span>
<span class="k">pass</span>
<span class="kn">import</span> <span class="nn">os</span>
<span class="kn">import</span> <span class="nn">tempfile</span>
<span class="kn">from</span> <span class="nn">collections</span> <span class="kn">import</span> <span class="n">defaultdict</span>
<span class="kn">from</span> <span class="nn">nltk.classify.api</span> <span class="kn">import</span> <span class="n">ClassifierI</span>
<span class="kn">from</span> <span class="nn">nltk.classify.megam</span> <span class="kn">import</span> <span class="n">call_megam</span><span class="p">,</span> <span class="n">parse_megam_weights</span><span class="p">,</span> <span class="n">write_megam_file</span>
<span class="kn">from</span> <span class="nn">nltk.classify.tadm</span> <span class="kn">import</span> <span class="n">call_tadm</span><span class="p">,</span> <span class="n">parse_tadm_weights</span><span class="p">,</span> <span class="n">write_tadm_file</span>
<span class="kn">from</span> <span class="nn">nltk.classify.util</span> <span class="kn">import</span> <span class="n">CutoffChecker</span><span class="p">,</span> <span class="n">accuracy</span><span class="p">,</span> <span class="n">log_likelihood</span>
<span class="kn">from</span> <span class="nn">nltk.data</span> <span class="kn">import</span> <span class="n">gzip_open_unicode</span>
<span class="kn">from</span> <span class="nn">nltk.probability</span> <span class="kn">import</span> <span class="n">DictionaryProbDist</span>
<span class="kn">from</span> <span class="nn">nltk.util</span> <span class="kn">import</span> <span class="n">OrderedDict</span>
<span class="n">__docformat__</span> <span class="o">=</span> <span class="s2">"epytext en"</span>
<span class="c1">######################################################################</span>
<span class="c1"># { Classifier Model</span>
<span class="c1">######################################################################</span>
<div class="viewcode-block" id="MaxentClassifier"><a class="viewcode-back" href="../../../api/nltk.classify.maxent.html#nltk.classify.maxent.MaxentClassifier">[docs]</a><span class="k">class</span> <span class="nc">MaxentClassifier</span><span class="p">(</span><span class="n">ClassifierI</span><span class="p">):</span>
<span class="sd">"""</span>
<span class="sd"> A maximum entropy classifier (also known as a "conditional</span>
<span class="sd"> exponential classifier"). This classifier is parameterized by a</span>
<span class="sd"> set of "weights", which are used to combine the joint-features</span>
<span class="sd"> that are generated from a featureset by an "encoding". In</span>
<span class="sd"> particular, the encoding maps each ``(featureset, label)`` pair to</span>
<span class="sd"> a vector. The probability of each label is then computed using</span>
<span class="sd"> the following equation::</span>
<span class="sd"> dotprod(weights, encode(fs,label))</span>
<span class="sd"> prob(fs|label) = ---------------------------------------------------</span>
<span class="sd"> sum(dotprod(weights, encode(fs,l)) for l in labels)</span>
<span class="sd"> Where ``dotprod`` is the dot product::</span>
<span class="sd"> dotprod(a,b) = sum(x*y for (x,y) in zip(a,b))</span>
<span class="sd"> """</span>
<div class="viewcode-block" id="MaxentClassifier.__init__"><a class="viewcode-back" href="../../../api/nltk.classify.maxent.html#nltk.classify.maxent.MaxentClassifier.__init__">[docs]</a> <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">encoding</span><span class="p">,</span> <span class="n">weights</span><span class="p">,</span> <span class="n">logarithmic</span><span class="o">=</span><span class="kc">True</span><span class="p">):</span>
<span class="sd">"""</span>
<span class="sd"> Construct a new maxent classifier model. Typically, new</span>
<span class="sd"> classifier models are created using the ``train()`` method.</span>
<span class="sd"> :type encoding: MaxentFeatureEncodingI</span>
<span class="sd"> :param encoding: An encoding that is used to convert the</span>
<span class="sd"> featuresets that are given to the ``classify`` method into</span>
<span class="sd"> joint-feature vectors, which are used by the maxent</span>
<span class="sd"> classifier model.</span>
<span class="sd"> :type weights: list of float</span>
<span class="sd"> :param weights: The feature weight vector for this classifier.</span>
<span class="sd"> :type logarithmic: bool</span>
<span class="sd"> :param logarithmic: If false, then use non-logarithmic weights.</span>
<span class="sd"> """</span>
<span class="bp">self</span><span class="o">.</span><span class="n">_encoding</span> <span class="o">=</span> <span class="n">encoding</span>
<span class="bp">self</span><span class="o">.</span><span class="n">_weights</span> <span class="o">=</span> <span class="n">weights</span>
<span class="bp">self</span><span class="o">.</span><span class="n">_logarithmic</span> <span class="o">=</span> <span class="n">logarithmic</span>
<span class="c1"># self._logarithmic = False</span>
<span class="k">assert</span> <span class="n">encoding</span><span class="o">.</span><span class="n">length</span><span class="p">()</span> <span class="o">==</span> <span class="nb">len</span><span class="p">(</span><span class="n">weights</span><span class="p">)</span></div>
<div class="viewcode-block" id="MaxentClassifier.labels"><a class="viewcode-back" href="../../../api/nltk.classify.maxent.html#nltk.classify.maxent.MaxentClassifier.labels">[docs]</a> <span class="k">def</span> <span class="nf">labels</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
<span class="k">return</span> <span class="bp">self</span><span class="o">.</span><span class="n">_encoding</span><span class="o">.</span><span class="n">labels</span><span class="p">()</span></div>
<div class="viewcode-block" id="MaxentClassifier.set_weights"><a class="viewcode-back" href="../../../api/nltk.classify.maxent.html#nltk.classify.maxent.MaxentClassifier.set_weights">[docs]</a> <span class="k">def</span> <span class="nf">set_weights</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">new_weights</span><span class="p">):</span>
<span class="sd">"""</span>
<span class="sd"> Set the feature weight vector for this classifier.</span>
<span class="sd"> :param new_weights: The new feature weight vector.</span>
<span class="sd"> :type new_weights: list of float</span>
<span class="sd"> """</span>
<span class="bp">self</span><span class="o">.</span><span class="n">_weights</span> <span class="o">=</span> <span class="n">new_weights</span>
<span class="k">assert</span> <span class="bp">self</span><span class="o">.</span><span class="n">_encoding</span><span class="o">.</span><span class="n">length</span><span class="p">()</span> <span class="o">==</span> <span class="nb">len</span><span class="p">(</span><span class="n">new_weights</span><span class="p">)</span></div>
<div class="viewcode-block" id="MaxentClassifier.weights"><a class="viewcode-back" href="../../../api/nltk.classify.maxent.html#nltk.classify.maxent.MaxentClassifier.weights">[docs]</a> <span class="k">def</span> <span class="nf">weights</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
<span class="sd">"""</span>
<span class="sd"> :return: The feature weight vector for this classifier.</span>
<span class="sd"> :rtype: list of float</span>
<span class="sd"> """</span>
<span class="k">return</span> <span class="bp">self</span><span class="o">.</span><span class="n">_weights</span></div>
<div class="viewcode-block" id="MaxentClassifier.classify"><a class="viewcode-back" href="../../../api/nltk.classify.maxent.html#nltk.classify.maxent.MaxentClassifier.classify">[docs]</a> <span class="k">def</span> <span class="nf">classify</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">featureset</span><span class="p">):</span>
<span class="k">return</span> <span class="bp">self</span><span class="o">.</span><span class="n">prob_classify</span><span class="p">(</span><span class="n">featureset</span><span class="p">)</span><span class="o">.</span><span class="n">max</span><span class="p">()</span></div>
<div class="viewcode-block" id="MaxentClassifier.prob_classify"><a class="viewcode-back" href="../../../api/nltk.classify.maxent.html#nltk.classify.maxent.MaxentClassifier.prob_classify">[docs]</a> <span class="k">def</span> <span class="nf">prob_classify</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">featureset</span><span class="p">):</span>
<span class="n">prob_dict</span> <span class="o">=</span> <span class="p">{}</span>
<span class="k">for</span> <span class="n">label</span> <span class="ow">in</span> <span class="bp">self</span><span class="o">.</span><span class="n">_encoding</span><span class="o">.</span><span class="n">labels</span><span class="p">():</span>
<span class="n">feature_vector</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">_encoding</span><span class="o">.</span><span class="n">encode</span><span class="p">(</span><span class="n">featureset</span><span class="p">,</span> <span class="n">label</span><span class="p">)</span>
<span class="k">if</span> <span class="bp">self</span><span class="o">.</span><span class="n">_logarithmic</span><span class="p">:</span>
<span class="n">total</span> <span class="o">=</span> <span class="mf">0.0</span>
<span class="k">for</span> <span class="p">(</span><span class="n">f_id</span><span class="p">,</span> <span class="n">f_val</span><span class="p">)</span> <span class="ow">in</span> <span class="n">feature_vector</span><span class="p">:</span>
<span class="n">total</span> <span class="o">+=</span> <span class="bp">self</span><span class="o">.</span><span class="n">_weights</span><span class="p">[</span><span class="n">f_id</span><span class="p">]</span> <span class="o">*</span> <span class="n">f_val</span>
<span class="n">prob_dict</span><span class="p">[</span><span class="n">label</span><span class="p">]</span> <span class="o">=</span> <span class="n">total</span>
<span class="k">else</span><span class="p">:</span>
<span class="n">prod</span> <span class="o">=</span> <span class="mf">1.0</span>
<span class="k">for</span> <span class="p">(</span><span class="n">f_id</span><span class="p">,</span> <span class="n">f_val</span><span class="p">)</span> <span class="ow">in</span> <span class="n">feature_vector</span><span class="p">:</span>
<span class="n">prod</span> <span class="o">*=</span> <span class="bp">self</span><span class="o">.</span><span class="n">_weights</span><span class="p">[</span><span class="n">f_id</span><span class="p">]</span> <span class="o">**</span> <span class="n">f_val</span>
<span class="n">prob_dict</span><span class="p">[</span><span class="n">label</span><span class="p">]</span> <span class="o">=</span> <span class="n">prod</span>
<span class="c1"># Normalize the dictionary to give a probability distribution</span>
<span class="k">return</span> <span class="n">DictionaryProbDist</span><span class="p">(</span><span class="n">prob_dict</span><span class="p">,</span> <span class="n">log</span><span class="o">=</span><span class="bp">self</span><span class="o">.</span><span class="n">_logarithmic</span><span class="p">,</span> <span class="n">normalize</span><span class="o">=</span><span class="kc">True</span><span class="p">)</span></div>
<div class="viewcode-block" id="MaxentClassifier.explain"><a class="viewcode-back" href="../../../api/nltk.classify.maxent.html#nltk.classify.maxent.MaxentClassifier.explain">[docs]</a> <span class="k">def</span> <span class="nf">explain</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">featureset</span><span class="p">,</span> <span class="n">columns</span><span class="o">=</span><span class="mi">4</span><span class="p">):</span>
<span class="sd">"""</span>
<span class="sd"> Print a table showing the effect of each of the features in</span>
<span class="sd"> the given feature set, and how they combine to determine the</span>
<span class="sd"> probabilities of each label for that featureset.</span>
<span class="sd"> """</span>
<span class="n">descr_width</span> <span class="o">=</span> <span class="mi">50</span>
<span class="n">TEMPLATE</span> <span class="o">=</span> <span class="s2">" %-"</span> <span class="o">+</span> <span class="nb">str</span><span class="p">(</span><span class="n">descr_width</span> <span class="o">-</span> <span class="mi">2</span><span class="p">)</span> <span class="o">+</span> <span class="s2">"s</span><span class="si">%s%8.3f</span><span class="s2">"</span>
<span class="n">pdist</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">prob_classify</span><span class="p">(</span><span class="n">featureset</span><span class="p">)</span>
<span class="n">labels</span> <span class="o">=</span> <span class="nb">sorted</span><span class="p">(</span><span class="n">pdist</span><span class="o">.</span><span class="n">samples</span><span class="p">(),</span> <span class="n">key</span><span class="o">=</span><span class="n">pdist</span><span class="o">.</span><span class="n">prob</span><span class="p">,</span> <span class="n">reverse</span><span class="o">=</span><span class="kc">True</span><span class="p">)</span>
<span class="n">labels</span> <span class="o">=</span> <span class="n">labels</span><span class="p">[:</span><span class="n">columns</span><span class="p">]</span>
<span class="nb">print</span><span class="p">(</span>
<span class="s2">" Feature"</span><span class="o">.</span><span class="n">ljust</span><span class="p">(</span><span class="n">descr_width</span><span class="p">)</span>
<span class="o">+</span> <span class="s2">""</span><span class="o">.</span><span class="n">join</span><span class="p">(</span><span class="s2">"</span><span class="si">%8s</span><span class="s2">"</span> <span class="o">%</span> <span class="p">((</span><span class="s2">"</span><span class="si">%s</span><span class="s2">"</span> <span class="o">%</span> <span class="n">l</span><span class="p">)[:</span><span class="mi">7</span><span class="p">])</span> <span class="k">for</span> <span class="n">l</span> <span class="ow">in</span> <span class="n">labels</span><span class="p">)</span>
<span class="p">)</span>
<span class="nb">print</span><span class="p">(</span><span class="s2">" "</span> <span class="o">+</span> <span class="s2">"-"</span> <span class="o">*</span> <span class="p">(</span><span class="n">descr_width</span> <span class="o">-</span> <span class="mi">2</span> <span class="o">+</span> <span class="mi">8</span> <span class="o">*</span> <span class="nb">len</span><span class="p">(</span><span class="n">labels</span><span class="p">)))</span>
<span class="n">sums</span> <span class="o">=</span> <span class="n">defaultdict</span><span class="p">(</span><span class="nb">int</span><span class="p">)</span>
<span class="k">for</span> <span class="n">i</span><span class="p">,</span> <span class="n">label</span> <span class="ow">in</span> <span class="nb">enumerate</span><span class="p">(</span><span class="n">labels</span><span class="p">):</span>
<span class="n">feature_vector</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">_encoding</span><span class="o">.</span><span class="n">encode</span><span class="p">(</span><span class="n">featureset</span><span class="p">,</span> <span class="n">label</span><span class="p">)</span>
<span class="n">feature_vector</span><span class="o">.</span><span class="n">sort</span><span class="p">(</span>
<span class="n">key</span><span class="o">=</span><span class="k">lambda</span> <span class="n">fid__</span><span class="p">:</span> <span class="nb">abs</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">_weights</span><span class="p">[</span><span class="n">fid__</span><span class="p">[</span><span class="mi">0</span><span class="p">]]),</span> <span class="n">reverse</span><span class="o">=</span><span class="kc">True</span>
<span class="p">)</span>
<span class="k">for</span> <span class="p">(</span><span class="n">f_id</span><span class="p">,</span> <span class="n">f_val</span><span class="p">)</span> <span class="ow">in</span> <span class="n">feature_vector</span><span class="p">:</span>
<span class="k">if</span> <span class="bp">self</span><span class="o">.</span><span class="n">_logarithmic</span><span class="p">:</span>
<span class="n">score</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">_weights</span><span class="p">[</span><span class="n">f_id</span><span class="p">]</span> <span class="o">*</span> <span class="n">f_val</span>
<span class="k">else</span><span class="p">:</span>
<span class="n">score</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">_weights</span><span class="p">[</span><span class="n">f_id</span><span class="p">]</span> <span class="o">**</span> <span class="n">f_val</span>
<span class="n">descr</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">_encoding</span><span class="o">.</span><span class="n">describe</span><span class="p">(</span><span class="n">f_id</span><span class="p">)</span>
<span class="n">descr</span> <span class="o">=</span> <span class="n">descr</span><span class="o">.</span><span class="n">split</span><span class="p">(</span><span class="s2">" and label is "</span><span class="p">)[</span><span class="mi">0</span><span class="p">]</span> <span class="c1"># hack</span>
<span class="n">descr</span> <span class="o">+=</span> <span class="s2">" (</span><span class="si">%s</span><span class="s2">)"</span> <span class="o">%</span> <span class="n">f_val</span> <span class="c1"># hack</span>
<span class="k">if</span> <span class="nb">len</span><span class="p">(</span><span class="n">descr</span><span class="p">)</span> <span class="o">></span> <span class="mi">47</span><span class="p">:</span>
<span class="n">descr</span> <span class="o">=</span> <span class="n">descr</span><span class="p">[:</span><span class="mi">44</span><span class="p">]</span> <span class="o">+</span> <span class="s2">"..."</span>
<span class="nb">print</span><span class="p">(</span><span class="n">TEMPLATE</span> <span class="o">%</span> <span class="p">(</span><span class="n">descr</span><span class="p">,</span> <span class="n">i</span> <span class="o">*</span> <span class="mi">8</span> <span class="o">*</span> <span class="s2">" "</span><span class="p">,</span> <span class="n">score</span><span class="p">))</span>
<span class="n">sums</span><span class="p">[</span><span class="n">label</span><span class="p">]</span> <span class="o">+=</span> <span class="n">score</span>
<span class="nb">print</span><span class="p">(</span><span class="s2">" "</span> <span class="o">+</span> <span class="s2">"-"</span> <span class="o">*</span> <span class="p">(</span><span class="n">descr_width</span> <span class="o">-</span> <span class="mi">1</span> <span class="o">+</span> <span class="mi">8</span> <span class="o">*</span> <span class="nb">len</span><span class="p">(</span><span class="n">labels</span><span class="p">)))</span>
<span class="nb">print</span><span class="p">(</span>
<span class="s2">" TOTAL:"</span><span class="o">.</span><span class="n">ljust</span><span class="p">(</span><span class="n">descr_width</span><span class="p">)</span> <span class="o">+</span> <span class="s2">""</span><span class="o">.</span><span class="n">join</span><span class="p">(</span><span class="s2">"</span><span class="si">%8.3f</span><span class="s2">"</span> <span class="o">%</span> <span class="n">sums</span><span class="p">[</span><span class="n">l</span><span class="p">]</span> <span class="k">for</span> <span class="n">l</span> <span class="ow">in</span> <span class="n">labels</span><span class="p">)</span>
<span class="p">)</span>
<span class="nb">print</span><span class="p">(</span>
<span class="s2">" PROBS:"</span><span class="o">.</span><span class="n">ljust</span><span class="p">(</span><span class="n">descr_width</span><span class="p">)</span>
<span class="o">+</span> <span class="s2">""</span><span class="o">.</span><span class="n">join</span><span class="p">(</span><span class="s2">"</span><span class="si">%8.3f</span><span class="s2">"</span> <span class="o">%</span> <span class="n">pdist</span><span class="o">.</span><span class="n">prob</span><span class="p">(</span><span class="n">l</span><span class="p">)</span> <span class="k">for</span> <span class="n">l</span> <span class="ow">in</span> <span class="n">labels</span><span class="p">)</span>
<span class="p">)</span></div>
<div class="viewcode-block" id="MaxentClassifier.most_informative_features"><a class="viewcode-back" href="../../../api/nltk.classify.maxent.html#nltk.classify.maxent.MaxentClassifier.most_informative_features">[docs]</a> <span class="k">def</span> <span class="nf">most_informative_features</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">n</span><span class="o">=</span><span class="mi">10</span><span class="p">):</span>
<span class="sd">"""</span>
<span class="sd"> Generates the ranked list of informative features from most to least.</span>
<span class="sd"> """</span>
<span class="k">if</span> <span class="nb">hasattr</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="s2">"_most_informative_features"</span><span class="p">):</span>
<span class="k">return</span> <span class="bp">self</span><span class="o">.</span><span class="n">_most_informative_features</span><span class="p">[:</span><span class="n">n</span><span class="p">]</span>
<span class="k">else</span><span class="p">:</span>
<span class="bp">self</span><span class="o">.</span><span class="n">_most_informative_features</span> <span class="o">=</span> <span class="nb">sorted</span><span class="p">(</span>
<span class="nb">list</span><span class="p">(</span><span class="nb">range</span><span class="p">(</span><span class="nb">len</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">_weights</span><span class="p">))),</span>
<span class="n">key</span><span class="o">=</span><span class="k">lambda</span> <span class="n">fid</span><span class="p">:</span> <span class="nb">abs</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">_weights</span><span class="p">[</span><span class="n">fid</span><span class="p">]),</span>
<span class="n">reverse</span><span class="o">=</span><span class="kc">True</span><span class="p">,</span>
<span class="p">)</span>
<span class="k">return</span> <span class="bp">self</span><span class="o">.</span><span class="n">_most_informative_features</span><span class="p">[:</span><span class="n">n</span><span class="p">]</span></div>
<div class="viewcode-block" id="MaxentClassifier.show_most_informative_features"><a class="viewcode-back" href="../../../api/nltk.classify.maxent.html#nltk.classify.maxent.MaxentClassifier.show_most_informative_features">[docs]</a> <span class="k">def</span> <span class="nf">show_most_informative_features</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">n</span><span class="o">=</span><span class="mi">10</span><span class="p">,</span> <span class="n">show</span><span class="o">=</span><span class="s2">"all"</span><span class="p">):</span>
<span class="sd">"""</span>
<span class="sd"> :param show: all, neg, or pos (for negative-only or positive-only)</span>
<span class="sd"> :type show: str</span>
<span class="sd"> :param n: The no. of top features</span>
<span class="sd"> :type n: int</span>
<span class="sd"> """</span>
<span class="c1"># Use None the full list of ranked features.</span>
<span class="n">fids</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">most_informative_features</span><span class="p">(</span><span class="kc">None</span><span class="p">)</span>
<span class="k">if</span> <span class="n">show</span> <span class="o">==</span> <span class="s2">"pos"</span><span class="p">:</span>
<span class="n">fids</span> <span class="o">=</span> <span class="p">[</span><span class="n">fid</span> <span class="k">for</span> <span class="n">fid</span> <span class="ow">in</span> <span class="n">fids</span> <span class="k">if</span> <span class="bp">self</span><span class="o">.</span><span class="n">_weights</span><span class="p">[</span><span class="n">fid</span><span class="p">]</span> <span class="o">></span> <span class="mi">0</span><span class="p">]</span>
<span class="k">elif</span> <span class="n">show</span> <span class="o">==</span> <span class="s2">"neg"</span><span class="p">:</span>
<span class="n">fids</span> <span class="o">=</span> <span class="p">[</span><span class="n">fid</span> <span class="k">for</span> <span class="n">fid</span> <span class="ow">in</span> <span class="n">fids</span> <span class="k">if</span> <span class="bp">self</span><span class="o">.</span><span class="n">_weights</span><span class="p">[</span><span class="n">fid</span><span class="p">]</span> <span class="o"><</span> <span class="mi">0</span><span class="p">]</span>
<span class="k">for</span> <span class="n">fid</span> <span class="ow">in</span> <span class="n">fids</span><span class="p">[:</span><span class="n">n</span><span class="p">]:</span>
<span class="nb">print</span><span class="p">(</span><span class="sa">f</span><span class="s2">"</span><span class="si">{</span><span class="bp">self</span><span class="o">.</span><span class="n">_weights</span><span class="p">[</span><span class="n">fid</span><span class="p">]</span><span class="si">:</span><span class="s2">8.3f</span><span class="si">}</span><span class="s2"> </span><span class="si">{</span><span class="bp">self</span><span class="o">.</span><span class="n">_encoding</span><span class="o">.</span><span class="n">describe</span><span class="p">(</span><span class="n">fid</span><span class="p">)</span><span class="si">}</span><span class="s2">"</span><span class="p">)</span></div>
<span class="k">def</span> <span class="fm">__repr__</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
<span class="k">return</span> <span class="s2">"<ConditionalExponentialClassifier: </span><span class="si">%d</span><span class="s2"> labels, </span><span class="si">%d</span><span class="s2"> features>"</span> <span class="o">%</span> <span class="p">(</span>
<span class="nb">len</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">_encoding</span><span class="o">.</span><span class="n">labels</span><span class="p">()),</span>
<span class="bp">self</span><span class="o">.</span><span class="n">_encoding</span><span class="o">.</span><span class="n">length</span><span class="p">(),</span>
<span class="p">)</span>
<span class="c1">#: A list of the algorithm names that are accepted for the</span>
<span class="c1">#: ``train()`` method's ``algorithm`` parameter.</span>
<span class="n">ALGORITHMS</span> <span class="o">=</span> <span class="p">[</span><span class="s2">"GIS"</span><span class="p">,</span> <span class="s2">"IIS"</span><span class="p">,</span> <span class="s2">"MEGAM"</span><span class="p">,</span> <span class="s2">"TADM"</span><span class="p">]</span>
<div class="viewcode-block" id="MaxentClassifier.train"><a class="viewcode-back" href="../../../api/nltk.classify.maxent.html#nltk.classify.maxent.MaxentClassifier.train">[docs]</a> <span class="nd">@classmethod</span>
<span class="k">def</span> <span class="nf">train</span><span class="p">(</span>
<span class="bp">cls</span><span class="p">,</span>
<span class="n">train_toks</span><span class="p">,</span>
<span class="n">algorithm</span><span class="o">=</span><span class="kc">None</span><span class="p">,</span>
<span class="n">trace</span><span class="o">=</span><span class="mi">3</span><span class="p">,</span>
<span class="n">encoding</span><span class="o">=</span><span class="kc">None</span><span class="p">,</span>
<span class="n">labels</span><span class="o">=</span><span class="kc">None</span><span class="p">,</span>
<span class="n">gaussian_prior_sigma</span><span class="o">=</span><span class="mi">0</span><span class="p">,</span>
<span class="o">**</span><span class="n">cutoffs</span><span class="p">,</span>
<span class="p">):</span>
<span class="sd">"""</span>
<span class="sd"> Train a new maxent classifier based on the given corpus of</span>
<span class="sd"> training samples. This classifier will have its weights</span>
<span class="sd"> chosen to maximize entropy while remaining empirically</span>
<span class="sd"> consistent with the training corpus.</span>
<span class="sd"> :rtype: MaxentClassifier</span>
<span class="sd"> :return: The new maxent classifier</span>
<span class="sd"> :type train_toks: list</span>
<span class="sd"> :param train_toks: Training data, represented as a list of</span>
<span class="sd"> pairs, the first member of which is a featureset,</span>
<span class="sd"> and the second of which is a classification label.</span>
<span class="sd"> :type algorithm: str</span>
<span class="sd"> :param algorithm: A case-insensitive string, specifying which</span>
<span class="sd"> algorithm should be used to train the classifier. The</span>
<span class="sd"> following algorithms are currently available.</span>
<span class="sd"> - Iterative Scaling Methods: Generalized Iterative Scaling (``'GIS'``),</span>
<span class="sd"> Improved Iterative Scaling (``'IIS'``)</span>
<span class="sd"> - External Libraries (requiring megam):</span>
<span class="sd"> LM-BFGS algorithm, with training performed by Megam (``'megam'``)</span>
<span class="sd"> The default algorithm is ``'IIS'``.</span>
<span class="sd"> :type trace: int</span>
<span class="sd"> :param trace: The level of diagnostic tracing output to produce.</span>
<span class="sd"> Higher values produce more verbose output.</span>
<span class="sd"> :type encoding: MaxentFeatureEncodingI</span>
<span class="sd"> :param encoding: A feature encoding, used to convert featuresets</span>
<span class="sd"> into feature vectors. If none is specified, then a</span>
<span class="sd"> ``BinaryMaxentFeatureEncoding`` will be built based on the</span>
<span class="sd"> features that are attested in the training corpus.</span>
<span class="sd"> :type labels: list(str)</span>
<span class="sd"> :param labels: The set of possible labels. If none is given, then</span>
<span class="sd"> the set of all labels attested in the training data will be</span>
<span class="sd"> used instead.</span>
<span class="sd"> :param gaussian_prior_sigma: The sigma value for a gaussian</span>
<span class="sd"> prior on model weights. Currently, this is supported by</span>
<span class="sd"> ``megam``. For other algorithms, its value is ignored.</span>
<span class="sd"> :param cutoffs: Arguments specifying various conditions under</span>
<span class="sd"> which the training should be halted. (Some of the cutoff</span>
<span class="sd"> conditions are not supported by some algorithms.)</span>
<span class="sd"> - ``max_iter=v``: Terminate after ``v`` iterations.</span>
<span class="sd"> - ``min_ll=v``: Terminate after the negative average</span>
<span class="sd"> log-likelihood drops under ``v``.</span>
<span class="sd"> - ``min_lldelta=v``: Terminate if a single iteration improves</span>
<span class="sd"> log likelihood by less than ``v``.</span>
<span class="sd"> """</span>
<span class="k">if</span> <span class="n">algorithm</span> <span class="ow">is</span> <span class="kc">None</span><span class="p">:</span>
<span class="n">algorithm</span> <span class="o">=</span> <span class="s2">"iis"</span>
<span class="k">for</span> <span class="n">key</span> <span class="ow">in</span> <span class="n">cutoffs</span><span class="p">:</span>
<span class="k">if</span> <span class="n">key</span> <span class="ow">not</span> <span class="ow">in</span> <span class="p">(</span>
<span class="s2">"max_iter"</span><span class="p">,</span>
<span class="s2">"min_ll"</span><span class="p">,</span>
<span class="s2">"min_lldelta"</span><span class="p">,</span>
<span class="s2">"max_acc"</span><span class="p">,</span>
<span class="s2">"min_accdelta"</span><span class="p">,</span>
<span class="s2">"count_cutoff"</span><span class="p">,</span>
<span class="s2">"norm"</span><span class="p">,</span>
<span class="s2">"explicit"</span><span class="p">,</span>
<span class="s2">"bernoulli"</span><span class="p">,</span>
<span class="p">):</span>
<span class="k">raise</span> <span class="ne">TypeError</span><span class="p">(</span><span class="s2">"Unexpected keyword arg </span><span class="si">%r</span><span class="s2">"</span> <span class="o">%</span> <span class="n">key</span><span class="p">)</span>
<span class="n">algorithm</span> <span class="o">=</span> <span class="n">algorithm</span><span class="o">.</span><span class="n">lower</span><span class="p">()</span>
<span class="k">if</span> <span class="n">algorithm</span> <span class="o">==</span> <span class="s2">"iis"</span><span class="p">:</span>
<span class="k">return</span> <span class="n">train_maxent_classifier_with_iis</span><span class="p">(</span>
<span class="n">train_toks</span><span class="p">,</span> <span class="n">trace</span><span class="p">,</span> <span class="n">encoding</span><span class="p">,</span> <span class="n">labels</span><span class="p">,</span> <span class="o">**</span><span class="n">cutoffs</span>
<span class="p">)</span>
<span class="k">elif</span> <span class="n">algorithm</span> <span class="o">==</span> <span class="s2">"gis"</span><span class="p">:</span>
<span class="k">return</span> <span class="n">train_maxent_classifier_with_gis</span><span class="p">(</span>
<span class="n">train_toks</span><span class="p">,</span> <span class="n">trace</span><span class="p">,</span> <span class="n">encoding</span><span class="p">,</span> <span class="n">labels</span><span class="p">,</span> <span class="o">**</span><span class="n">cutoffs</span>
<span class="p">)</span>
<span class="k">elif</span> <span class="n">algorithm</span> <span class="o">==</span> <span class="s2">"megam"</span><span class="p">:</span>
<span class="k">return</span> <span class="n">train_maxent_classifier_with_megam</span><span class="p">(</span>
<span class="n">train_toks</span><span class="p">,</span> <span class="n">trace</span><span class="p">,</span> <span class="n">encoding</span><span class="p">,</span> <span class="n">labels</span><span class="p">,</span> <span class="n">gaussian_prior_sigma</span><span class="p">,</span> <span class="o">**</span><span class="n">cutoffs</span>
<span class="p">)</span>
<span class="k">elif</span> <span class="n">algorithm</span> <span class="o">==</span> <span class="s2">"tadm"</span><span class="p">:</span>
<span class="n">kwargs</span> <span class="o">=</span> <span class="n">cutoffs</span>
<span class="n">kwargs</span><span class="p">[</span><span class="s2">"trace"</span><span class="p">]</span> <span class="o">=</span> <span class="n">trace</span>
<span class="n">kwargs</span><span class="p">[</span><span class="s2">"encoding"</span><span class="p">]</span> <span class="o">=</span> <span class="n">encoding</span>
<span class="n">kwargs</span><span class="p">[</span><span class="s2">"labels"</span><span class="p">]</span> <span class="o">=</span> <span class="n">labels</span>
<span class="n">kwargs</span><span class="p">[</span><span class="s2">"gaussian_prior_sigma"</span><span class="p">]</span> <span class="o">=</span> <span class="n">gaussian_prior_sigma</span>
<span class="k">return</span> <span class="n">TadmMaxentClassifier</span><span class="o">.</span><span class="n">train</span><span class="p">(</span><span class="n">train_toks</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">)</span>
<span class="k">else</span><span class="p">:</span>
<span class="k">raise</span> <span class="ne">ValueError</span><span class="p">(</span><span class="s2">"Unknown algorithm </span><span class="si">%s</span><span class="s2">"</span> <span class="o">%</span> <span class="n">algorithm</span><span class="p">)</span></div></div>
<span class="c1">#: Alias for MaxentClassifier.</span>
<span class="n">ConditionalExponentialClassifier</span> <span class="o">=</span> <span class="n">MaxentClassifier</span>
<span class="c1">######################################################################</span>
<span class="c1"># { Feature Encodings</span>
<span class="c1">######################################################################</span>
<div class="viewcode-block" id="MaxentFeatureEncodingI"><a class="viewcode-back" href="../../../api/nltk.classify.maxent.html#nltk.classify.maxent.MaxentFeatureEncodingI">[docs]</a><span class="k">class</span> <span class="nc">MaxentFeatureEncodingI</span><span class="p">:</span>
<span class="sd">"""</span>
<span class="sd"> A mapping that converts a set of input-feature values to a vector</span>
<span class="sd"> of joint-feature values, given a label. This conversion is</span>
<span class="sd"> necessary to translate featuresets into a format that can be used</span>
<span class="sd"> by maximum entropy models.</span>
<span class="sd"> The set of joint-features used by a given encoding is fixed, and</span>
<span class="sd"> each index in the generated joint-feature vectors corresponds to a</span>
<span class="sd"> single joint-feature. The length of the generated joint-feature</span>
<span class="sd"> vectors is therefore constant (for a given encoding).</span>
<span class="sd"> Because the joint-feature vectors generated by</span>
<span class="sd"> ``MaxentFeatureEncodingI`` are typically very sparse, they are</span>
<span class="sd"> represented as a list of ``(index, value)`` tuples, specifying the</span>
<span class="sd"> value of each non-zero joint-feature.</span>
<span class="sd"> Feature encodings are generally created using the ``train()``</span>
<span class="sd"> method, which generates an appropriate encoding based on the</span>
<span class="sd"> input-feature values and labels that are present in a given</span>
<span class="sd"> corpus.</span>
<span class="sd"> """</span>
<div class="viewcode-block" id="MaxentFeatureEncodingI.encode"><a class="viewcode-back" href="../../../api/nltk.classify.maxent.html#nltk.classify.maxent.MaxentFeatureEncodingI.encode">[docs]</a> <span class="k">def</span> <span class="nf">encode</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">featureset</span><span class="p">,</span> <span class="n">label</span><span class="p">):</span>
<span class="sd">"""</span>
<span class="sd"> Given a (featureset, label) pair, return the corresponding</span>
<span class="sd"> vector of joint-feature values. This vector is represented as</span>
<span class="sd"> a list of ``(index, value)`` tuples, specifying the value of</span>
<span class="sd"> each non-zero joint-feature.</span>
<span class="sd"> :type featureset: dict</span>
<span class="sd"> :rtype: list(tuple(int, int))</span>
<span class="sd"> """</span>
<span class="k">raise</span> <span class="ne">NotImplementedError</span><span class="p">()</span></div>
<div class="viewcode-block" id="MaxentFeatureEncodingI.length"><a class="viewcode-back" href="../../../api/nltk.classify.maxent.html#nltk.classify.maxent.MaxentFeatureEncodingI.length">[docs]</a> <span class="k">def</span> <span class="nf">length</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
<span class="sd">"""</span>
<span class="sd"> :return: The size of the fixed-length joint-feature vectors</span>
<span class="sd"> that are generated by this encoding.</span>
<span class="sd"> :rtype: int</span>
<span class="sd"> """</span>
<span class="k">raise</span> <span class="ne">NotImplementedError</span><span class="p">()</span></div>
<div class="viewcode-block" id="MaxentFeatureEncodingI.labels"><a class="viewcode-back" href="../../../api/nltk.classify.maxent.html#nltk.classify.maxent.MaxentFeatureEncodingI.labels">[docs]</a> <span class="k">def</span> <span class="nf">labels</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
<span class="sd">"""</span>
<span class="sd"> :return: A list of the \"known labels\" -- i.e., all labels</span>
<span class="sd"> ``l`` such that ``self.encode(fs,l)`` can be a nonzero</span>
<span class="sd"> joint-feature vector for some value of ``fs``.</span>
<span class="sd"> :rtype: list</span>
<span class="sd"> """</span>
<span class="k">raise</span> <span class="ne">NotImplementedError</span><span class="p">()</span></div>
<div class="viewcode-block" id="MaxentFeatureEncodingI.describe"><a class="viewcode-back" href="../../../api/nltk.classify.maxent.html#nltk.classify.maxent.MaxentFeatureEncodingI.describe">[docs]</a> <span class="k">def</span> <span class="nf">describe</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">fid</span><span class="p">):</span>
<span class="sd">"""</span>
<span class="sd"> :return: A string describing the value of the joint-feature</span>
<span class="sd"> whose index in the generated feature vectors is ``fid``.</span>
<span class="sd"> :rtype: str</span>
<span class="sd"> """</span>
<span class="k">raise</span> <span class="ne">NotImplementedError</span><span class="p">()</span></div>
<div class="viewcode-block" id="MaxentFeatureEncodingI.train"><a class="viewcode-back" href="../../../api/nltk.classify.maxent.html#nltk.classify.maxent.MaxentFeatureEncodingI.train">[docs]</a> <span class="k">def</span> <span class="nf">train</span><span class="p">(</span><span class="bp">cls</span><span class="p">,</span> <span class="n">train_toks</span><span class="p">):</span>
<span class="sd">"""</span>
<span class="sd"> Construct and return new feature encoding, based on a given</span>
<span class="sd"> training corpus ``train_toks``.</span>
<span class="sd"> :type train_toks: list(tuple(dict, str))</span>
<span class="sd"> :param train_toks: Training data, represented as a list of</span>
<span class="sd"> pairs, the first member of which is a feature dictionary,</span>
<span class="sd"> and the second of which is a classification label.</span>
<span class="sd"> """</span>
<span class="k">raise</span> <span class="ne">NotImplementedError</span><span class="p">()</span></div></div>
<div class="viewcode-block" id="FunctionBackedMaxentFeatureEncoding"><a class="viewcode-back" href="../../../api/nltk.classify.maxent.html#nltk.classify.maxent.FunctionBackedMaxentFeatureEncoding">[docs]</a><span class="k">class</span> <span class="nc">FunctionBackedMaxentFeatureEncoding</span><span class="p">(</span><span class="n">MaxentFeatureEncodingI</span><span class="p">):</span>
<span class="sd">"""</span>
<span class="sd"> A feature encoding that calls a user-supplied function to map a</span>
<span class="sd"> given featureset/label pair to a sparse joint-feature vector.</span>
<span class="sd"> """</span>
<div class="viewcode-block" id="FunctionBackedMaxentFeatureEncoding.__init__"><a class="viewcode-back" href="../../../api/nltk.classify.maxent.html#nltk.classify.maxent.FunctionBackedMaxentFeatureEncoding.__init__">[docs]</a> <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">func</span><span class="p">,</span> <span class="n">length</span><span class="p">,</span> <span class="n">labels</span><span class="p">):</span>
<span class="sd">"""</span>
<span class="sd"> Construct a new feature encoding based on the given function.</span>
<span class="sd"> :type func: (callable)</span>
<span class="sd"> :param func: A function that takes two arguments, a featureset</span>
<span class="sd"> and a label, and returns the sparse joint feature vector</span>
<span class="sd"> that encodes them::</span>
<span class="sd"> func(featureset, label) -> feature_vector</span>
<span class="sd"> This sparse joint feature vector (``feature_vector``) is a</span>
<span class="sd"> list of ``(index,value)`` tuples.</span>
<span class="sd"> :type length: int</span>
<span class="sd"> :param length: The size of the fixed-length joint-feature</span>
<span class="sd"> vectors that are generated by this encoding.</span>
<span class="sd"> :type labels: list</span>
<span class="sd"> :param labels: A list of the \"known labels\" for this</span>
<span class="sd"> encoding -- i.e., all labels ``l`` such that</span>
<span class="sd"> ``self.encode(fs,l)`` can be a nonzero joint-feature vector</span>
<span class="sd"> for some value of ``fs``.</span>
<span class="sd"> """</span>
<span class="bp">self</span><span class="o">.</span><span class="n">_length</span> <span class="o">=</span> <span class="n">length</span>
<span class="bp">self</span><span class="o">.</span><span class="n">_func</span> <span class="o">=</span> <span class="n">func</span>
<span class="bp">self</span><span class="o">.</span><span class="n">_labels</span> <span class="o">=</span> <span class="n">labels</span></div>
<div class="viewcode-block" id="FunctionBackedMaxentFeatureEncoding.encode"><a class="viewcode-back" href="../../../api/nltk.classify.maxent.html#nltk.classify.maxent.FunctionBackedMaxentFeatureEncoding.encode">[docs]</a> <span class="k">def</span> <span class="nf">encode</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">featureset</span><span class="p">,</span> <span class="n">label</span><span class="p">):</span>
<span class="k">return</span> <span class="bp">self</span><span class="o">.</span><span class="n">_func</span><span class="p">(</span><span class="n">featureset</span><span class="p">,</span> <span class="n">label</span><span class="p">)</span></div>
<div class="viewcode-block" id="FunctionBackedMaxentFeatureEncoding.length"><a class="viewcode-back" href="../../../api/nltk.classify.maxent.html#nltk.classify.maxent.FunctionBackedMaxentFeatureEncoding.length">[docs]</a> <span class="k">def</span> <span class="nf">length</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
<span class="k">return</span> <span class="bp">self</span><span class="o">.</span><span class="n">_length</span></div>
<div class="viewcode-block" id="FunctionBackedMaxentFeatureEncoding.labels"><a class="viewcode-back" href="../../../api/nltk.classify.maxent.html#nltk.classify.maxent.FunctionBackedMaxentFeatureEncoding.labels">[docs]</a> <span class="k">def</span> <span class="nf">labels</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
<span class="k">return</span> <span class="bp">self</span><span class="o">.</span><span class="n">_labels</span></div>
<div class="viewcode-block" id="FunctionBackedMaxentFeatureEncoding.describe"><a class="viewcode-back" href="../../../api/nltk.classify.maxent.html#nltk.classify.maxent.FunctionBackedMaxentFeatureEncoding.describe">[docs]</a> <span class="k">def</span> <span class="nf">describe</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">fid</span><span class="p">):</span>
<span class="k">return</span> <span class="s2">"no description available"</span></div></div>
<div class="viewcode-block" id="BinaryMaxentFeatureEncoding"><a class="viewcode-back" href="../../../api/nltk.classify.maxent.html#nltk.classify.maxent.BinaryMaxentFeatureEncoding">[docs]</a><span class="k">class</span> <span class="nc">BinaryMaxentFeatureEncoding</span><span class="p">(</span><span class="n">MaxentFeatureEncodingI</span><span class="p">):</span>
<span class="sd">"""</span>
<span class="sd"> A feature encoding that generates vectors containing a binary</span>
<span class="sd"> joint-features of the form:</span>
<span class="sd"> | joint_feat(fs, l) = { 1 if (fs[fname] == fval) and (l == label)</span>
<span class="sd"> | {</span>
<span class="sd"> | { 0 otherwise</span>
<span class="sd"> Where ``fname`` is the name of an input-feature, ``fval`` is a value</span>
<span class="sd"> for that input-feature, and ``label`` is a label.</span>
<span class="sd"> Typically, these features are constructed based on a training</span>
<span class="sd"> corpus, using the ``train()`` method. This method will create one</span>
<span class="sd"> feature for each combination of ``fname``, ``fval``, and ``label``</span>
<span class="sd"> that occurs at least once in the training corpus.</span>
<span class="sd"> The ``unseen_features`` parameter can be used to add "unseen-value</span>
<span class="sd"> features", which are used whenever an input feature has a value</span>
<span class="sd"> that was not encountered in the training corpus. These features</span>
<span class="sd"> have the form:</span>
<span class="sd"> | joint_feat(fs, l) = { 1 if is_unseen(fname, fs[fname])</span>
<span class="sd"> | { and l == label</span>
<span class="sd"> | {</span>
<span class="sd"> | { 0 otherwise</span>
<span class="sd"> Where ``is_unseen(fname, fval)`` is true if the encoding does not</span>
<span class="sd"> contain any joint features that are true when ``fs[fname]==fval``.</span>
<span class="sd"> The ``alwayson_features`` parameter can be used to add "always-on</span>
<span class="sd"> features", which have the form::</span>
<span class="sd"> | joint_feat(fs, l) = { 1 if (l == label)</span>
<span class="sd"> | {</span>
<span class="sd"> | { 0 otherwise</span>
<span class="sd"> These always-on features allow the maxent model to directly model</span>
<span class="sd"> the prior probabilities of each label.</span>
<span class="sd"> """</span>
<div class="viewcode-block" id="BinaryMaxentFeatureEncoding.__init__"><a class="viewcode-back" href="../../../api/nltk.classify.maxent.html#nltk.classify.maxent.BinaryMaxentFeatureEncoding.__init__">[docs]</a> <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">labels</span><span class="p">,</span> <span class="n">mapping</span><span class="p">,</span> <span class="n">unseen_features</span><span class="o">=</span><span class="kc">False</span><span class="p">,</span> <span class="n">alwayson_features</span><span class="o">=</span><span class="kc">False</span><span class="p">):</span>
<span class="sd">"""</span>
<span class="sd"> :param labels: A list of the \"known labels\" for this encoding.</span>
<span class="sd"> :param mapping: A dictionary mapping from ``(fname,fval,label)``</span>
<span class="sd"> tuples to corresponding joint-feature indexes. These</span>
<span class="sd"> indexes must be the set of integers from 0...len(mapping).</span>
<span class="sd"> If ``mapping[fname,fval,label]=id``, then</span>
<span class="sd"> ``self.encode(..., fname:fval, ..., label)[id]`` is 1;</span>
<span class="sd"> otherwise, it is 0.</span>
<span class="sd"> :param unseen_features: If true, then include unseen value</span>
<span class="sd"> features in the generated joint-feature vectors.</span>
<span class="sd"> :param alwayson_features: If true, then include always-on</span>
<span class="sd"> features in the generated joint-feature vectors.</span>
<span class="sd"> """</span>
<span class="k">if</span> <span class="nb">set</span><span class="p">(</span><span class="n">mapping</span><span class="o">.</span><span class="n">values</span><span class="p">())</span> <span class="o">!=</span> <span class="nb">set</span><span class="p">(</span><span class="nb">range</span><span class="p">(</span><span class="nb">len</span><span class="p">(</span><span class="n">mapping</span><span class="p">))):</span>
<span class="k">raise</span> <span class="ne">ValueError</span><span class="p">(</span>
<span class="s2">"Mapping values must be exactly the "</span>
<span class="s2">"set of integers from 0...len(mapping)"</span>
<span class="p">)</span>
<span class="bp">self</span><span class="o">.</span><span class="n">_labels</span> <span class="o">=</span> <span class="nb">list</span><span class="p">(</span><span class="n">labels</span><span class="p">)</span>
<span class="sd">"""A list of attested labels."""</span>
<span class="bp">self</span><span class="o">.</span><span class="n">_mapping</span> <span class="o">=</span> <span class="n">mapping</span>
<span class="sd">"""dict mapping from (fname,fval,label) -> fid"""</span>
<span class="bp">self</span><span class="o">.</span><span class="n">_length</span> <span class="o">=</span> <span class="nb">len</span><span class="p">(</span><span class="n">mapping</span><span class="p">)</span>
<span class="sd">"""The length of generated joint feature vectors."""</span>
<span class="bp">self</span><span class="o">.</span><span class="n">_alwayson</span> <span class="o">=</span> <span class="kc">None</span>
<span class="sd">"""dict mapping from label -> fid"""</span>
<span class="bp">self</span><span class="o">.</span><span class="n">_unseen</span> <span class="o">=</span> <span class="kc">None</span>
<span class="sd">"""dict mapping from fname -> fid"""</span>
<span class="k">if</span> <span class="n">alwayson_features</span><span class="p">:</span>
<span class="bp">self</span><span class="o">.</span><span class="n">_alwayson</span> <span class="o">=</span> <span class="p">{</span>
<span class="n">label</span><span class="p">:</span> <span class="n">i</span> <span class="o">+</span> <span class="bp">self</span><span class="o">.</span><span class="n">_length</span> <span class="k">for</span> <span class="p">(</span><span class="n">i</span><span class="p">,</span> <span class="n">label</span><span class="p">)</span> <span class="ow">in</span> <span class="nb">enumerate</span><span class="p">(</span><span class="n">labels</span><span class="p">)</span>
<span class="p">}</span>
<span class="bp">self</span><span class="o">.</span><span class="n">_length</span> <span class="o">+=</span> <span class="nb">len</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">_alwayson</span><span class="p">)</span>
<span class="k">if</span> <span class="n">unseen_features</span><span class="p">:</span>
<span class="n">fnames</span> <span class="o">=</span> <span class="p">{</span><span class="n">fname</span> <span class="k">for</span> <span class="p">(</span><span class="n">fname</span><span class="p">,</span> <span class="n">fval</span><span class="p">,</span> <span class="n">label</span><span class="p">)</span> <span class="ow">in</span> <span class="n">mapping</span><span class="p">}</span>
<span class="bp">self</span><span class="o">.</span><span class="n">_unseen</span> <span class="o">=</span> <span class="p">{</span><span class="n">fname</span><span class="p">:</span> <span class="n">i</span> <span class="o">+</span> <span class="bp">self</span><span class="o">.</span><span class="n">_length</span> <span class="k">for</span> <span class="p">(</span><span class="n">i</span><span class="p">,</span> <span class="n">fname</span><span class="p">)</span> <span class="ow">in</span> <span class="nb">enumerate</span><span class="p">(</span><span class="n">fnames</span><span class="p">)}</span>
<span class="bp">self</span><span class="o">.</span><span class="n">_length</span> <span class="o">+=</span> <span class="nb">len</span><span class="p">(</span><span class="n">fnames</span><span class="p">)</span></div>
<div class="viewcode-block" id="BinaryMaxentFeatureEncoding.encode"><a class="viewcode-back" href="../../../api/nltk.classify.maxent.html#nltk.classify.maxent.BinaryMaxentFeatureEncoding.encode">[docs]</a> <span class="k">def</span> <span class="nf">encode</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">featureset</span><span class="p">,</span> <span class="n">label</span><span class="p">):</span>
<span class="c1"># Inherit docs.</span>
<span class="n">encoding</span> <span class="o">=</span> <span class="p">[]</span>
<span class="c1"># Convert input-features to joint-features:</span>
<span class="k">for</span> <span class="n">fname</span><span class="p">,</span> <span class="n">fval</span> <span class="ow">in</span> <span class="n">featureset</span><span class="o">.</span><span class="n">items</span><span class="p">():</span>
<span class="c1"># Known feature name & value:</span>
<span class="k">if</span> <span class="p">(</span><span class="n">fname</span><span class="p">,</span> <span class="n">fval</span><span class="p">,</span> <span class="n">label</span><span class="p">)</span> <span class="ow">in</span> <span class="bp">self</span><span class="o">.</span><span class="n">_mapping</span><span class="p">:</span>
<span class="n">encoding</span><span class="o">.</span><span class="n">append</span><span class="p">((</span><span class="bp">self</span><span class="o">.</span><span class="n">_mapping</span><span class="p">[</span><span class="n">fname</span><span class="p">,</span> <span class="n">fval</span><span class="p">,</span> <span class="n">label</span><span class="p">],</span> <span class="mi">1</span><span class="p">))</span>
<span class="c1"># Otherwise, we might want to fire an "unseen-value feature".</span>
<span class="k">elif</span> <span class="bp">self</span><span class="o">.</span><span class="n">_unseen</span><span class="p">:</span>
<span class="c1"># Have we seen this fname/fval combination with any label?</span>
<span class="k">for</span> <span class="n">label2</span> <span class="ow">in</span> <span class="bp">self</span><span class="o">.</span><span class="n">_labels</span><span class="p">:</span>
<span class="k">if</span> <span class="p">(</span><span class="n">fname</span><span class="p">,</span> <span class="n">fval</span><span class="p">,</span> <span class="n">label2</span><span class="p">)</span> <span class="ow">in</span> <span class="bp">self</span><span class="o">.</span><span class="n">_mapping</span><span class="p">:</span>
<span class="k">break</span> <span class="c1"># we've seen this fname/fval combo</span>
<span class="c1"># We haven't -- fire the unseen-value feature</span>
<span class="k">else</span><span class="p">:</span>
<span class="k">if</span> <span class="n">fname</span> <span class="ow">in</span> <span class="bp">self</span><span class="o">.</span><span class="n">_unseen</span><span class="p">:</span>
<span class="n">encoding</span><span class="o">.</span><span class="n">append</span><span class="p">((</span><span class="bp">self</span><span class="o">.</span><span class="n">_unseen</span><span class="p">[</span><span class="n">fname</span><span class="p">],</span> <span class="mi">1</span><span class="p">))</span>
<span class="c1"># Add always-on features:</span>
<span class="k">if</span> <span class="bp">self</span><span class="o">.</span><span class="n">_alwayson</span> <span class="ow">and</span> <span class="n">label</span> <span class="ow">in</span> <span class="bp">self</span><span class="o">.</span><span class="n">_alwayson</span><span class="p">:</span>
<span class="n">encoding</span><span class="o">.</span><span class="n">append</span><span class="p">((</span><span class="bp">self</span><span class="o">.</span><span class="n">_alwayson</span><span class="p">[</span><span class="n">label</span><span class="p">],</span> <span class="mi">1</span><span class="p">))</span>
<span class="k">return</span> <span class="n">encoding</span></div>
<div class="viewcode-block" id="BinaryMaxentFeatureEncoding.describe"><a class="viewcode-back" href="../../../api/nltk.classify.maxent.html#nltk.classify.maxent.BinaryMaxentFeatureEncoding.describe">[docs]</a> <span class="k">def</span> <span class="nf">describe</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">f_id</span><span class="p">):</span>
<span class="c1"># Inherit docs.</span>
<span class="k">if</span> <span class="ow">not</span> <span class="nb">isinstance</span><span class="p">(</span><span class="n">f_id</span><span class="p">,</span> <span class="nb">int</span><span class="p">):</span>
<span class="k">raise</span> <span class="ne">TypeError</span><span class="p">(</span><span class="s2">"describe() expected an int"</span><span class="p">)</span>
<span class="k">try</span><span class="p">:</span>
<span class="bp">self</span><span class="o">.</span><span class="n">_inv_mapping</span>
<span class="k">except</span> <span class="ne">AttributeError</span><span class="p">:</span>
<span class="bp">self</span><span class="o">.</span><span class="n">_inv_mapping</span> <span class="o">=</span> <span class="p">[</span><span class="o">-</span><span class="mi">1</span><span class="p">]</span> <span class="o">*</span> <span class="nb">len</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">_mapping</span><span class="p">)</span>
<span class="k">for</span> <span class="p">(</span><span class="n">info</span><span class="p">,</span> <span class="n">i</span><span class="p">)</span> <span class="ow">in</span> <span class="bp">self</span><span class="o">.</span><span class="n">_mapping</span><span class="o">.</span><span class="n">items</span><span class="p">():</span>
<span class="bp">self</span><span class="o">.</span><span class="n">_inv_mapping</span><span class="p">[</span><span class="n">i</span><span class="p">]</span> <span class="o">=</span> <span class="n">info</span>
<span class="k">if</span> <span class="n">f_id</span> <span class="o"><</span> <span class="nb">len</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">_mapping</span><span class="p">):</span>
<span class="p">(</span><span class="n">fname</span><span class="p">,</span> <span class="n">fval</span><span class="p">,</span> <span class="n">label</span><span class="p">)</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">_inv_mapping</span><span class="p">[</span><span class="n">f_id</span><span class="p">]</span>
<span class="k">return</span> <span class="sa">f</span><span class="s2">"</span><span class="si">{</span><span class="n">fname</span><span class="si">}</span><span class="s2">==</span><span class="si">{</span><span class="n">fval</span><span class="si">!r}</span><span class="s2"> and label is </span><span class="si">{</span><span class="n">label</span><span class="si">!r}</span><span class="s2">"</span>
<span class="k">elif</span> <span class="bp">self</span><span class="o">.</span><span class="n">_alwayson</span> <span class="ow">and</span> <span class="n">f_id</span> <span class="ow">in</span> <span class="bp">self</span><span class="o">.</span><span class="n">_alwayson</span><span class="o">.</span><span class="n">values</span><span class="p">():</span>
<span class="k">for</span> <span class="p">(</span><span class="n">label</span><span class="p">,</span> <span class="n">f_id2</span><span class="p">)</span> <span class="ow">in</span> <span class="bp">self</span><span class="o">.</span><span class="n">_alwayson</span><span class="o">.</span><span class="n">items</span><span class="p">():</span>
<span class="k">if</span> <span class="n">f_id</span> <span class="o">==</span> <span class="n">f_id2</span><span class="p">:</span>
<span class="k">return</span> <span class="s2">"label is </span><span class="si">%r</span><span class="s2">"</span> <span class="o">%</span> <span class="n">label</span>
<span class="k">elif</span> <span class="bp">self</span><span class="o">.</span><span class="n">_unseen</span> <span class="ow">and</span> <span class="n">f_id</span> <span class="ow">in</span> <span class="bp">self</span><span class="o">.</span><span class="n">_unseen</span><span class="o">.</span><span class="n">values</span><span class="p">():</span>
<span class="k">for</span> <span class="p">(</span><span class="n">fname</span><span class="p">,</span> <span class="n">f_id2</span><span class="p">)</span> <span class="ow">in</span> <span class="bp">self</span><span class="o">.</span><span class="n">_unseen</span><span class="o">.</span><span class="n">items</span><span class="p">():</span>
<span class="k">if</span> <span class="n">f_id</span> <span class="o">==</span> <span class="n">f_id2</span><span class="p">:</span>
<span class="k">return</span> <span class="s2">"</span><span class="si">%s</span><span class="s2"> is unseen"</span> <span class="o">%</span> <span class="n">fname</span>
<span class="k">else</span><span class="p">:</span>
<span class="k">raise</span> <span class="ne">ValueError</span><span class="p">(</span><span class="s2">"Bad feature id"</span><span class="p">)</span></div>
<div class="viewcode-block" id="BinaryMaxentFeatureEncoding.labels"><a class="viewcode-back" href="../../../api/nltk.classify.maxent.html#nltk.classify.maxent.BinaryMaxentFeatureEncoding.labels">[docs]</a> <span class="k">def</span> <span class="nf">labels</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
<span class="c1"># Inherit docs.</span>
<span class="k">return</span> <span class="bp">self</span><span class="o">.</span><span class="n">_labels</span></div>
<div class="viewcode-block" id="BinaryMaxentFeatureEncoding.length"><a class="viewcode-back" href="../../../api/nltk.classify.maxent.html#nltk.classify.maxent.BinaryMaxentFeatureEncoding.length">[docs]</a> <span class="k">def</span> <span class="nf">length</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
<span class="c1"># Inherit docs.</span>
<span class="k">return</span> <span class="bp">self</span><span class="o">.</span><span class="n">_length</span></div>
<div class="viewcode-block" id="BinaryMaxentFeatureEncoding.train"><a class="viewcode-back" href="../../../api/nltk.classify.maxent.html#nltk.classify.maxent.BinaryMaxentFeatureEncoding.train">[docs]</a> <span class="nd">@classmethod</span>
<span class="k">def</span> <span class="nf">train</span><span class="p">(</span><span class="bp">cls</span><span class="p">,</span> <span class="n">train_toks</span><span class="p">,</span> <span class="n">count_cutoff</span><span class="o">=</span><span class="mi">0</span><span class="p">,</span> <span class="n">labels</span><span class="o">=</span><span class="kc">None</span><span class="p">,</span> <span class="o">**</span><span class="n">options</span><span class="p">):</span>
<span class="sd">"""</span>
<span class="sd"> Construct and return new feature encoding, based on a given</span>
<span class="sd"> training corpus ``train_toks``. See the class description</span>
<span class="sd"> ``BinaryMaxentFeatureEncoding`` for a description of the</span>
<span class="sd"> joint-features that will be included in this encoding.</span>
<span class="sd"> :type train_toks: list(tuple(dict, str))</span>
<span class="sd"> :param train_toks: Training data, represented as a list of</span>
<span class="sd"> pairs, the first member of which is a feature dictionary,</span>
<span class="sd"> and the second of which is a classification label.</span>
<span class="sd"> :type count_cutoff: int</span>
<span class="sd"> :param count_cutoff: A cutoff value that is used to discard</span>
<span class="sd"> rare joint-features. If a joint-feature's value is 1</span>
<span class="sd"> fewer than ``count_cutoff`` times in the training corpus,</span>
<span class="sd"> then that joint-feature is not included in the generated</span>
<span class="sd"> encoding.</span>
<span class="sd"> :type labels: list</span>
<span class="sd"> :param labels: A list of labels that should be used by the</span>
<span class="sd"> classifier. If not specified, then the set of labels</span>
<span class="sd"> attested in ``train_toks`` will be used.</span>
<span class="sd"> :param options: Extra parameters for the constructor, such as</span>
<span class="sd"> ``unseen_features`` and ``alwayson_features``.</span>
<span class="sd"> """</span>
<span class="n">mapping</span> <span class="o">=</span> <span class="p">{}</span> <span class="c1"># maps (fname, fval, label) -> fid</span>
<span class="n">seen_labels</span> <span class="o">=</span> <span class="nb">set</span><span class="p">()</span> <span class="c1"># The set of labels we've encountered</span>
<span class="n">count</span> <span class="o">=</span> <span class="n">defaultdict</span><span class="p">(</span><span class="nb">int</span><span class="p">)</span> <span class="c1"># maps (fname, fval) -> count</span>
<span class="k">for</span> <span class="p">(</span><span class="n">tok</span><span class="p">,</span> <span class="n">label</span><span class="p">)</span> <span class="ow">in</span> <span class="n">train_toks</span><span class="p">:</span>
<span class="k">if</span> <span class="n">labels</span> <span class="ow">and</span> <span class="n">label</span> <span class="ow">not</span> <span class="ow">in</span> <span class="n">labels</span><span class="p">:</span>
<span class="k">raise</span> <span class="ne">ValueError</span><span class="p">(</span><span class="s2">"Unexpected label </span><span class="si">%s</span><span class="s2">"</span> <span class="o">%</span> <span class="n">label</span><span class="p">)</span>
<span class="n">seen_labels</span><span class="o">.</span><span class="n">add</span><span class="p">(</span><span class="n">label</span><span class="p">)</span>
<span class="c1"># Record each of the features.</span>
<span class="k">for</span> <span class="p">(</span><span class="n">fname</span><span class="p">,</span> <span class="n">fval</span><span class="p">)</span> <span class="ow">in</span> <span class="n">tok</span><span class="o">.</span><span class="n">items</span><span class="p">():</span>
<span class="c1"># If a count cutoff is given, then only add a joint</span>
<span class="c1"># feature once the corresponding (fname, fval, label)</span>
<span class="c1"># tuple exceeds that cutoff.</span>
<span class="n">count</span><span class="p">[</span><span class="n">fname</span><span class="p">,</span> <span class="n">fval</span><span class="p">]</span> <span class="o">+=</span> <span class="mi">1</span>
<span class="k">if</span> <span class="n">count</span><span class="p">[</span><span class="n">fname</span><span class="p">,</span> <span class="n">fval</span><span class="p">]</span> <span class="o">>=</span> <span class="n">count_cutoff</span><span class="p">:</span>
<span class="k">if</span> <span class="p">(</span><span class="n">fname</span><span class="p">,</span> <span class="n">fval</span><span class="p">,</span> <span class="n">label</span><span class="p">)</span> <span class="ow">not</span> <span class="ow">in</span> <span class="n">mapping</span><span class="p">:</span>
<span class="n">mapping</span><span class="p">[</span><span class="n">fname</span><span class="p">,</span> <span class="n">fval</span><span class="p">,</span> <span class="n">label</span><span class="p">]</span> <span class="o">=</span> <span class="nb">len</span><span class="p">(</span><span class="n">mapping</span><span class="p">)</span>
<span class="k">if</span> <span class="n">labels</span> <span class="ow">is</span> <span class="kc">None</span><span class="p">:</span>
<span class="n">labels</span> <span class="o">=</span> <span class="n">seen_labels</span>
<span class="k">return</span> <span class="bp">cls</span><span class="p">(</span><span class="n">labels</span><span class="p">,</span> <span class="n">mapping</span><span class="p">,</span> <span class="o">**</span><span class="n">options</span><span class="p">)</span></div></div>
<div class="viewcode-block" id="GISEncoding"><a class="viewcode-back" href="../../../api/nltk.classify.maxent.html#nltk.classify.maxent.GISEncoding">[docs]</a><span class="k">class</span> <span class="nc">GISEncoding</span><span class="p">(</span><span class="n">BinaryMaxentFeatureEncoding</span><span class="p">):</span>
<span class="sd">"""</span>
<span class="sd"> A binary feature encoding which adds one new joint-feature to the</span>
<span class="sd"> joint-features defined by ``BinaryMaxentFeatureEncoding``: a</span>
<span class="sd"> correction feature, whose value is chosen to ensure that the</span>
<span class="sd"> sparse vector always sums to a constant non-negative number. This</span>
<span class="sd"> new feature is used to ensure two preconditions for the GIS</span>
<span class="sd"> training algorithm:</span>
<span class="sd"> - At least one feature vector index must be nonzero for every</span>
<span class="sd"> token.</span>
<span class="sd"> - The feature vector must sum to a constant non-negative number</span>
<span class="sd"> for every token.</span>
<span class="sd"> """</span>
<div class="viewcode-block" id="GISEncoding.__init__"><a class="viewcode-back" href="../../../api/nltk.classify.maxent.html#nltk.classify.maxent.GISEncoding.__init__">[docs]</a> <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span>
<span class="bp">self</span><span class="p">,</span> <span class="n">labels</span><span class="p">,</span> <span class="n">mapping</span><span class="p">,</span> <span class="n">unseen_features</span><span class="o">=</span><span class="kc">False</span><span class="p">,</span> <span class="n">alwayson_features</span><span class="o">=</span><span class="kc">False</span><span class="p">,</span> <span class="n">C</span><span class="o">=</span><span class="kc">None</span>
<span class="p">):</span>
<span class="sd">"""</span>
<span class="sd"> :param C: The correction constant. The value of the correction</span>
<span class="sd"> feature is based on this value. In particular, its value is</span>
<span class="sd"> ``C - sum([v for (f,v) in encoding])``.</span>
<span class="sd"> :seealso: ``BinaryMaxentFeatureEncoding.__init__``</span>
<span class="sd"> """</span>
<span class="n">BinaryMaxentFeatureEncoding</span><span class="o">.</span><span class="fm">__init__</span><span class="p">(</span>
<span class="bp">self</span><span class="p">,</span> <span class="n">labels</span><span class="p">,</span> <span class="n">mapping</span><span class="p">,</span> <span class="n">unseen_features</span><span class="p">,</span> <span class="n">alwayson_features</span>
<span class="p">)</span>
<span class="k">if</span> <span class="n">C</span> <span class="ow">is</span> <span class="kc">None</span><span class="p">:</span>
<span class="n">C</span> <span class="o">=</span> <span class="nb">len</span><span class="p">({</span><span class="n">fname</span> <span class="k">for</span> <span class="p">(</span><span class="n">fname</span><span class="p">,</span> <span class="n">fval</span><span class="p">,</span> <span class="n">label</span><span class="p">)</span> <span class="ow">in</span> <span class="n">mapping</span><span class="p">})</span> <span class="o">+</span> <span class="mi">1</span>
<span class="bp">self</span><span class="o">.</span><span class="n">_C</span> <span class="o">=</span> <span class="n">C</span></div>
<span class="nd">@property</span>
<span class="k">def</span> <span class="nf">C</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
<span class="sd">"""The non-negative constant that all encoded feature vectors</span>
<span class="sd"> will sum to."""</span>
<span class="k">return</span> <span class="bp">self</span><span class="o">.</span><span class="n">_C</span>
<div class="viewcode-block" id="GISEncoding.encode"><a class="viewcode-back" href="../../../api/nltk.classify.maxent.html#nltk.classify.maxent.GISEncoding.encode">[docs]</a> <span class="k">def</span> <span class="nf">encode</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">featureset</span><span class="p">,</span> <span class="n">label</span><span class="p">):</span>
<span class="c1"># Get the basic encoding.</span>
<span class="n">encoding</span> <span class="o">=</span> <span class="n">BinaryMaxentFeatureEncoding</span><span class="o">.</span><span class="n">encode</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">featureset</span><span class="p">,</span> <span class="n">label</span><span class="p">)</span>
<span class="n">base_length</span> <span class="o">=</span> <span class="n">BinaryMaxentFeatureEncoding</span><span class="o">.</span><span class="n">length</span><span class="p">(</span><span class="bp">self</span><span class="p">)</span>
<span class="c1"># Add a correction feature.</span>
<span class="n">total</span> <span class="o">=</span> <span class="nb">sum</span><span class="p">(</span><span class="n">v</span> <span class="k">for</span> <span class="p">(</span><span class="n">f</span><span class="p">,</span> <span class="n">v</span><span class="p">)</span> <span class="ow">in</span> <span class="n">encoding</span><span class="p">)</span>
<span class="k">if</span> <span class="n">total</span> <span class="o">>=</span> <span class="bp">self</span><span class="o">.</span><span class="n">_C</span><span class="p">:</span>
<span class="k">raise</span> <span class="ne">ValueError</span><span class="p">(</span><span class="s2">"Correction feature is not high enough!"</span><span class="p">)</span>
<span class="n">encoding</span><span class="o">.</span><span class="n">append</span><span class="p">((</span><span class="n">base_length</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">_C</span> <span class="o">-</span> <span class="n">total</span><span class="p">))</span>
<span class="c1"># Return the result</span>
<span class="k">return</span> <span class="n">encoding</span></div>
<div class="viewcode-block" id="GISEncoding.length"><a class="viewcode-back" href="../../../api/nltk.classify.maxent.html#nltk.classify.maxent.GISEncoding.length">[docs]</a> <span class="k">def</span> <span class="nf">length</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
<span class="k">return</span> <span class="n">BinaryMaxentFeatureEncoding</span><span class="o">.</span><span class="n">length</span><span class="p">(</span><span class="bp">self</span><span class="p">)</span> <span class="o">+</span> <span class="mi">1</span></div>
<div class="viewcode-block" id="GISEncoding.describe"><a class="viewcode-back" href="../../../api/nltk.classify.maxent.html#nltk.classify.maxent.GISEncoding.describe">[docs]</a> <span class="k">def</span> <span class="nf">describe</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">f_id</span><span class="p">):</span>
<span class="k">if</span> <span class="n">f_id</span> <span class="o">==</span> <span class="n">BinaryMaxentFeatureEncoding</span><span class="o">.</span><span class="n">length</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
<span class="k">return</span> <span class="s2">"Correction feature (</span><span class="si">%s</span><span class="s2">)"</span> <span class="o">%</span> <span class="bp">self</span><span class="o">.</span><span class="n">_C</span>
<span class="k">else</span><span class="p">:</span>
<span class="k">return</span> <span class="n">BinaryMaxentFeatureEncoding</span><span class="o">.</span><span class="n">describe</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">f_id</span><span class="p">)</span></div></div>
<div class="viewcode-block" id="TadmEventMaxentFeatureEncoding"><a class="viewcode-back" href="../../../api/nltk.classify.maxent.html#nltk.classify.maxent.TadmEventMaxentFeatureEncoding">[docs]</a><span class="k">class</span> <span class="nc">TadmEventMaxentFeatureEncoding</span><span class="p">(</span><span class="n">BinaryMaxentFeatureEncoding</span><span class="p">):</span>
<div class="viewcode-block" id="TadmEventMaxentFeatureEncoding.__init__"><a class="viewcode-back" href="../../../api/nltk.classify.maxent.html#nltk.classify.maxent.TadmEventMaxentFeatureEncoding.__init__">[docs]</a> <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">labels</span><span class="p">,</span> <span class="n">mapping</span><span class="p">,</span> <span class="n">unseen_features</span><span class="o">=</span><span class="kc">False</span><span class="p">,</span> <span class="n">alwayson_features</span><span class="o">=</span><span class="kc">False</span><span class="p">):</span>
<span class="bp">self</span><span class="o">.</span><span class="n">_mapping</span> <span class="o">=</span> <span class="n">OrderedDict</span><span class="p">(</span><span class="n">mapping</span><span class="p">)</span>
<span class="bp">self</span><span class="o">.</span><span class="n">_label_mapping</span> <span class="o">=</span> <span class="n">OrderedDict</span><span class="p">()</span>
<span class="n">BinaryMaxentFeatureEncoding</span><span class="o">.</span><span class="fm">__init__</span><span class="p">(</span>
<span class="bp">self</span><span class="p">,</span> <span class="n">labels</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">_mapping</span><span class="p">,</span> <span class="n">unseen_features</span><span class="p">,</span> <span class="n">alwayson_features</span>
<span class="p">)</span></div>
<div class="viewcode-block" id="TadmEventMaxentFeatureEncoding.encode"><a class="viewcode-back" href="../../../api/nltk.classify.maxent.html#nltk.classify.maxent.TadmEventMaxentFeatureEncoding.encode">[docs]</a> <span class="k">def</span> <span class="nf">encode</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">featureset</span><span class="p">,</span> <span class="n">label</span><span class="p">):</span>
<span class="n">encoding</span> <span class="o">=</span> <span class="p">[]</span>
<span class="k">for</span> <span class="n">feature</span><span class="p">,</span> <span class="n">value</span> <span class="ow">in</span> <span class="n">featureset</span><span class="o">.</span><span class="n">items</span><span class="p">():</span>
<span class="k">if</span> <span class="p">(</span><span class="n">feature</span><span class="p">,</span> <span class="n">label</span><span class="p">)</span> <span class="ow">not</span> <span class="ow">in</span> <span class="bp">self</span><span class="o">.</span><span class="n">_mapping</span><span class="p">:</span>
<span class="bp">self</span><span class="o">.</span><span class="n">_mapping</span><span class="p">[(</span><span class="n">feature</span><span class="p">,</span> <span class="n">label</span><span class="p">)]</span> <span class="o">=</span> <span class="nb">len</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">_mapping</span><span class="p">)</span>
<span class="k">if</span> <span class="n">value</span> <span class="ow">not</span> <span class="ow">in</span> <span class="bp">self</span><span class="o">.</span><span class="n">_label_mapping</span><span class="p">:</span>
<span class="k">if</span> <span class="ow">not</span> <span class="nb">isinstance</span><span class="p">(</span><span class="n">value</span><span class="p">,</span> <span class="nb">int</span><span class="p">):</span>
<span class="bp">self</span><span class="o">.</span><span class="n">_label_mapping</span><span class="p">[</span><span class="n">value</span><span class="p">]</span> <span class="o">=</span> <span class="nb">len</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">_label_mapping</span><span class="p">)</span>
<span class="k">else</span><span class="p">:</span>
<span class="bp">self</span><span class="o">.</span><span class="n">_label_mapping</span><span class="p">[</span><span class="n">value</span><span class="p">]</span> <span class="o">=</span> <span class="n">value</span>
<span class="n">encoding</span><span class="o">.</span><span class="n">append</span><span class="p">(</span>
<span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">_mapping</span><span class="p">[(</span><span class="n">feature</span><span class="p">,</span> <span class="n">label</span><span class="p">)],</span> <span class="bp">self</span><span class="o">.</span><span class="n">_label_mapping</span><span class="p">[</span><span class="n">value</span><span class="p">])</span>
<span class="p">)</span>
<span class="k">return</span> <span class="n">encoding</span></div>
<div class="viewcode-block" id="TadmEventMaxentFeatureEncoding.labels"><a class="viewcode-back" href="../../../api/nltk.classify.maxent.html#nltk.classify.maxent.TadmEventMaxentFeatureEncoding.labels">[docs]</a> <span class="k">def</span> <span class="nf">labels</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
<span class="k">return</span> <span class="bp">self</span><span class="o">.</span><span class="n">_labels</span></div>
<div class="viewcode-block" id="TadmEventMaxentFeatureEncoding.describe"><a class="viewcode-back" href="../../../api/nltk.classify.maxent.html#nltk.classify.maxent.TadmEventMaxentFeatureEncoding.describe">[docs]</a> <span class="k">def</span> <span class="nf">describe</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">fid</span><span class="p">):</span>
<span class="k">for</span> <span class="p">(</span><span class="n">feature</span><span class="p">,</span> <span class="n">label</span><span class="p">)</span> <span class="ow">in</span> <span class="bp">self</span><span class="o">.</span><span class="n">_mapping</span><span class="p">:</span>
<span class="k">if</span> <span class="bp">self</span><span class="o">.</span><span class="n">_mapping</span><span class="p">[(</span><span class="n">feature</span><span class="p">,</span> <span class="n">label</span><span class="p">)]</span> <span class="o">==</span> <span class="n">fid</span><span class="p">:</span>
<span class="k">return</span> <span class="p">(</span><span class="n">feature</span><span class="p">,</span> <span class="n">label</span><span class="p">)</span></div>
<div class="viewcode-block" id="TadmEventMaxentFeatureEncoding.length"><a class="viewcode-back" href="../../../api/nltk.classify.maxent.html#nltk.classify.maxent.TadmEventMaxentFeatureEncoding.length">[docs]</a> <span class="k">def</span> <span class="nf">length</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
<span class="k">return</span> <span class="nb">len</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">_mapping</span><span class="p">)</span></div>
<div class="viewcode-block" id="TadmEventMaxentFeatureEncoding.train"><a class="viewcode-back" href="../../../api/nltk.classify.maxent.html#nltk.classify.maxent.TadmEventMaxentFeatureEncoding.train">[docs]</a> <span class="nd">@classmethod</span>
<span class="k">def</span> <span class="nf">train</span><span class="p">(</span><span class="bp">cls</span><span class="p">,</span> <span class="n">train_toks</span><span class="p">,</span> <span class="n">count_cutoff</span><span class="o">=</span><span class="mi">0</span><span class="p">,</span> <span class="n">labels</span><span class="o">=</span><span class="kc">None</span><span class="p">,</span> <span class="o">**</span><span class="n">options</span><span class="p">):</span>
<span class="n">mapping</span> <span class="o">=</span> <span class="n">OrderedDict</span><span class="p">()</span>
<span class="k">if</span> <span class="ow">not</span> <span class="n">labels</span><span class="p">:</span>
<span class="n">labels</span> <span class="o">=</span> <span class="p">[]</span>
<span class="c1"># This gets read twice, so compute the values in case it's lazy.</span>
<span class="n">train_toks</span> <span class="o">=</span> <span class="nb">list</span><span class="p">(</span><span class="n">train_toks</span><span class="p">)</span>
<span class="k">for</span> <span class="p">(</span><span class="n">featureset</span><span class="p">,</span> <span class="n">label</span><span class="p">)</span> <span class="ow">in</span> <span class="n">train_toks</span><span class="p">:</span>
<span class="k">if</span> <span class="n">label</span> <span class="ow">not</span> <span class="ow">in</span> <span class="n">labels</span><span class="p">:</span>
<span class="n">labels</span><span class="o">.</span><span class="n">append</span><span class="p">(</span><span class="n">label</span><span class="p">)</span>
<span class="k">for</span> <span class="p">(</span><span class="n">featureset</span><span class="p">,</span> <span class="n">label</span><span class="p">)</span> <span class="ow">in</span> <span class="n">train_toks</span><span class="p">:</span>
<span class="k">for</span> <span class="n">label</span> <span class="ow">in</span> <span class="n">labels</span><span class="p">:</span>
<span class="k">for</span> <span class="n">feature</span> <span class="ow">in</span> <span class="n">featureset</span><span class="p">:</span>
<span class="k">if</span> <span class="p">(</span><span class="n">feature</span><span class="p">,</span> <span class="n">label</span><span class="p">)</span> <span class="ow">not</span> <span class="ow">in</span> <span class="n">mapping</span><span class="p">:</span>
<span class="n">mapping</span><span class="p">[(</span><span class="n">feature</span><span class="p">,</span> <span class="n">label</span><span class="p">)]</span> <span class="o">=</span> <span class="nb">len</span><span class="p">(</span><span class="n">mapping</span><span class="p">)</span>
<span class="k">return</span> <span class="bp">cls</span><span class="p">(</span><span class="n">labels</span><span class="p">,</span> <span class="n">mapping</span><span class="p">,</span> <span class="o">**</span><span class="n">options</span><span class="p">)</span></div></div>
<div class="viewcode-block" id="TypedMaxentFeatureEncoding"><a class="viewcode-back" href="../../../api/nltk.classify.maxent.html#nltk.classify.maxent.TypedMaxentFeatureEncoding">[docs]</a><span class="k">class</span> <span class="nc">TypedMaxentFeatureEncoding</span><span class="p">(</span><span class="n">MaxentFeatureEncodingI</span><span class="p">):</span>
<span class="sd">"""</span>
<span class="sd"> A feature encoding that generates vectors containing integer,</span>
<span class="sd"> float and binary joint-features of the form:</span>
<span class="sd"> Binary (for string and boolean features):</span>
<span class="sd"> | joint_feat(fs, l) = { 1 if (fs[fname] == fval) and (l == label)</span>
<span class="sd"> | {</span>
<span class="sd"> | { 0 otherwise</span>
<span class="sd"> Value (for integer and float features):</span>
<span class="sd"> | joint_feat(fs, l) = { fval if (fs[fname] == type(fval))</span>
<span class="sd"> | { and (l == label)</span>
<span class="sd"> | {</span>
<span class="sd"> | { not encoded otherwise</span>
<span class="sd"> Where ``fname`` is the name of an input-feature, ``fval`` is a value</span>
<span class="sd"> for that input-feature, and ``label`` is a label.</span>
<span class="sd"> Typically, these features are constructed based on a training</span>
<span class="sd"> corpus, using the ``train()`` method.</span>
<span class="sd"> For string and boolean features [type(fval) not in (int, float)]</span>
<span class="sd"> this method will create one feature for each combination of</span>
<span class="sd"> ``fname``, ``fval``, and ``label`` that occurs at least once in the</span>
<span class="sd"> training corpus.</span>
<span class="sd"> For integer and float features [type(fval) in (int, float)] this</span>
<span class="sd"> method will create one feature for each combination of ``fname``</span>
<span class="sd"> and ``label`` that occurs at least once in the training corpus.</span>
<span class="sd"> For binary features the ``unseen_features`` parameter can be used</span>
<span class="sd"> to add "unseen-value features", which are used whenever an input</span>
<span class="sd"> feature has a value that was not encountered in the training</span>
<span class="sd"> corpus. These features have the form:</span>
<span class="sd"> | joint_feat(fs, l) = { 1 if is_unseen(fname, fs[fname])</span>
<span class="sd"> | { and l == label</span>
<span class="sd"> | {</span>
<span class="sd"> | { 0 otherwise</span>
<span class="sd"> Where ``is_unseen(fname, fval)`` is true if the encoding does not</span>
<span class="sd"> contain any joint features that are true when ``fs[fname]==fval``.</span>
<span class="sd"> The ``alwayson_features`` parameter can be used to add "always-on</span>
<span class="sd"> features", which have the form:</span>
<span class="sd"> | joint_feat(fs, l) = { 1 if (l == label)</span>
<span class="sd"> | {</span>
<span class="sd"> | { 0 otherwise</span>
<span class="sd"> These always-on features allow the maxent model to directly model</span>
<span class="sd"> the prior probabilities of each label.</span>
<span class="sd"> """</span>
<div class="viewcode-block" id="TypedMaxentFeatureEncoding.__init__"><a class="viewcode-back" href="../../../api/nltk.classify.maxent.html#nltk.classify.maxent.TypedMaxentFeatureEncoding.__init__">[docs]</a> <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">labels</span><span class="p">,</span> <span class="n">mapping</span><span class="p">,</span> <span class="n">unseen_features</span><span class="o">=</span><span class="kc">False</span><span class="p">,</span> <span class="n">alwayson_features</span><span class="o">=</span><span class="kc">False</span><span class="p">):</span>
<span class="sd">"""</span>
<span class="sd"> :param labels: A list of the \"known labels\" for this encoding.</span>
<span class="sd"> :param mapping: A dictionary mapping from ``(fname,fval,label)``</span>
<span class="sd"> tuples to corresponding joint-feature indexes. These</span>
<span class="sd"> indexes must be the set of integers from 0...len(mapping).</span>
<span class="sd"> If ``mapping[fname,fval,label]=id``, then</span>
<span class="sd"> ``self.encode({..., fname:fval, ...``, label)[id]} is 1;</span>
<span class="sd"> otherwise, it is 0.</span>
<span class="sd"> :param unseen_features: If true, then include unseen value</span>
<span class="sd"> features in the generated joint-feature vectors.</span>
<span class="sd"> :param alwayson_features: If true, then include always-on</span>
<span class="sd"> features in the generated joint-feature vectors.</span>
<span class="sd"> """</span>
<span class="k">if</span> <span class="nb">set</span><span class="p">(</span><span class="n">mapping</span><span class="o">.</span><span class="n">values</span><span class="p">())</span> <span class="o">!=</span> <span class="nb">set</span><span class="p">(</span><span class="nb">range</span><span class="p">(</span><span class="nb">len</span><span class="p">(</span><span class="n">mapping</span><span class="p">))):</span>
<span class="k">raise</span> <span class="ne">ValueError</span><span class="p">(</span>
<span class="s2">"Mapping values must be exactly the "</span>
<span class="s2">"set of integers from 0...len(mapping)"</span>
<span class="p">)</span>
<span class="bp">self</span><span class="o">.</span><span class="n">_labels</span> <span class="o">=</span> <span class="nb">list</span><span class="p">(</span><span class="n">labels</span><span class="p">)</span>
<span class="sd">"""A list of attested labels."""</span>
<span class="bp">self</span><span class="o">.</span><span class="n">_mapping</span> <span class="o">=</span> <span class="n">mapping</span>
<span class="sd">"""dict mapping from (fname,fval,label) -> fid"""</span>
<span class="bp">self</span><span class="o">.</span><span class="n">_length</span> <span class="o">=</span> <span class="nb">len</span><span class="p">(</span><span class="n">mapping</span><span class="p">)</span>
<span class="sd">"""The length of generated joint feature vectors."""</span>
<span class="bp">self</span><span class="o">.</span><span class="n">_alwayson</span> <span class="o">=</span> <span class="kc">None</span>
<span class="sd">"""dict mapping from label -> fid"""</span>
<span class="bp">self</span><span class="o">.</span><span class="n">_unseen</span> <span class="o">=</span> <span class="kc">None</span>
<span class="sd">"""dict mapping from fname -> fid"""</span>
<span class="k">if</span> <span class="n">alwayson_features</span><span class="p">:</span>
<span class="bp">self</span><span class="o">.</span><span class="n">_alwayson</span> <span class="o">=</span> <span class="p">{</span>
<span class="n">label</span><span class="p">:</span> <span class="n">i</span> <span class="o">+</span> <span class="bp">self</span><span class="o">.</span><span class="n">_length</span> <span class="k">for</span> <span class="p">(</span><span class="n">i</span><span class="p">,</span> <span class="n">label</span><span class="p">)</span> <span class="ow">in</span> <span class="nb">enumerate</span><span class="p">(</span><span class="n">labels</span><span class="p">)</span>
<span class="p">}</span>
<span class="bp">self</span><span class="o">.</span><span class="n">_length</span> <span class="o">+=</span> <span class="nb">len</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">_alwayson</span><span class="p">)</span>
<span class="k">if</span> <span class="n">unseen_features</span><span class="p">:</span>
<span class="n">fnames</span> <span class="o">=</span> <span class="p">{</span><span class="n">fname</span> <span class="k">for</span> <span class="p">(</span><span class="n">fname</span><span class="p">,</span> <span class="n">fval</span><span class="p">,</span> <span class="n">label</span><span class="p">)</span> <span class="ow">in</span> <span class="n">mapping</span><span class="p">}</span>
<span class="bp">self</span><span class="o">.</span><span class="n">_unseen</span> <span class="o">=</span> <span class="p">{</span><span class="n">fname</span><span class="p">:</span> <span class="n">i</span> <span class="o">+</span> <span class="bp">self</span><span class="o">.</span><span class="n">_length</span> <span class="k">for</span> <span class="p">(</span><span class="n">i</span><span class="p">,</span> <span class="n">fname</span><span class="p">)</span> <span class="ow">in</span> <span class="nb">enumerate</span><span class="p">(</span><span class="n">fnames</span><span class="p">)}</span>
<span class="bp">self</span><span class="o">.</span><span class="n">_length</span> <span class="o">+=</span> <span class="nb">len</span><span class="p">(</span><span class="n">fnames</span><span class="p">)</span></div>
<div class="viewcode-block" id="TypedMaxentFeatureEncoding.encode"><a class="viewcode-back" href="../../../api/nltk.classify.maxent.html#nltk.classify.maxent.TypedMaxentFeatureEncoding.encode">[docs]</a> <span class="k">def</span> <span class="nf">encode</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">featureset</span><span class="p">,</span> <span class="n">label</span><span class="p">):</span>