teratail
回答率
85.35%
teratail header banner
teratail header banner
質問するログイン新規登録
トップに関する質問Random Forestのマルチラベル分類モデルについて

編集履歴

質問編集履歴

1

具体的なコードを入力

2019/11/02 13:11

投稿

natsunatsu
natsunatsu

スコア11

test CHANGED
File without changes
test CHANGED
@@ -12,7 +12,507 @@
12
12
 
13
13
  これはいったいどういうことなのでしょうか？
14
14
 
15
- ちなみにLogistic回帰やカーネルSVCなどで同じことをやると、きちんとpredict_probaで確率が算出されます。なぜかRandomForestClassifierだけうまくいきません。解決策が分かれば教えて頂けますと幸いです。
15
+ ちなみにLogistic回帰やカーネルSVCなどで同じことをやると、きちんとpredict_probaで確率が算出されます。なぜかRandomForestClassifierだけうまくいきません。解決策が分かれば教えて頂けますと幸いです。以下コードです。
16
+
17
+
18
+
19
+ import numpy as np
20
+
21
+ from pandas import Series,DataFrame
22
+
23
+ import pandas as pd
24
+
25
+ from sklearn.ensemble import RandomForestClassifier
26
+
27
+ from skmultilearn.problem_transform import BinaryRelevance
28
+
29
+
30
+
31
+ df1 = pd.read_excel('train.xlsx',sheetname='Sheet1')
32
+
33
+ df2 = pd.read_excel('test.xlsx',sheetname='Sheet1')
34
+
35
+
36
+
37
+ df1_x = df1.copy()
38
+
39
+ df1_y = df1.copy()
40
+
41
+
42
+
43
+ df1_x = df1_x.loc[:, "x1":"x777"]
44
+
45
+ df1_y = df1_y.loc[:, "a1":"a10"]
46
+
47
+ df1_x = df1_x.where(df1_x>0, 0)
48
+
49
+
50
+
51
+ >>> df1_y
52
+
53
+ a1 a2 a3 a4 a5 a6 a7 a8 a9 a10
54
+
55
+ 0 1 0 0 0 0 0 0 0 0 0
56
+
57
+ 1 1 0 0 0 0 0 0 0 0 0
58
+
59
+ 2 1 0 0 0 0 0 0 0 0 0
60
+
61
+ 3 1 0 0 0 0 0 0 0 0 0
62
+
63
+ 4 0 1 0 0 0 0 0 0 0 0
64
+
65
+ 5 0 1 0 0 0 0 0 0 0 0
66
+
67
+ 6 0 1 0 0 0 0 0 0 0 0
68
+
69
+ 7 0 1 0 0 0 0 0 0 0 0
70
+
71
+ 8 0 0 1 0 0 0 0 0 0 0
72
+
73
+ 9 0 0 1 0 0 0 0 0 0 0
74
+
75
+ 10 0 0 1 0 0 0 0 0 0 0
76
+
77
+ 11 0 0 1 0 0 0 0 0 0 0
78
+
79
+ 12 0 0 0 1 0 0 0 0 0 0
80
+
81
+ 13 0 0 0 1 0 0 0 0 0 0
82
+
83
+ 14 0 0 0 1 0 0 0 0 0 0
84
+
85
+ 15 0 0 0 1 0 0 0 0 0 0
86
+
87
+ 16 0 0 0 0 1 0 0 0 0 0
88
+
89
+ 17 0 0 0 0 1 0 0 0 0 0
90
+
91
+ 18 0 0 0 0 1 0 0 0 0 0
92
+
93
+ 19 0 0 0 0 1 0 0 0 0 0
94
+
95
+ 20 0 0 0 0 0 1 0 0 0 0
96
+
97
+ 21 0 0 0 0 0 1 0 0 0 0
98
+
99
+ 22 0 0 0 0 0 1 0 0 0 0
100
+
101
+ 23 0 0 0 0 0 1 0 0 0 0
102
+
103
+ 24 0 0 0 0 0 0 1 0 0 0
104
+
105
+ 25 0 0 0 0 0 0 1 0 0 0
106
+
107
+ 26 0 0 0 0 0 0 1 0 0 0
108
+
109
+ 27 0 0 0 0 0 0 1 0 0 0
110
+
111
+ 28 0 0 0 0 0 0 0 1 0 0
112
+
113
+ 29 0 0 0 0 0 0 0 1 0 0
114
+
115
+ 30 0 0 0 0 0 0 0 1 0 0
116
+
117
+ 31 0 0 0 0 0 0 0 1 0 0
118
+
119
+ 32 0 0 0 0 0 0 0 0 1 0
120
+
121
+ 33 0 0 0 0 0 0 0 0 1 0
122
+
123
+ 34 0 0 0 0 0 0 0 0 1 0
124
+
125
+ 35 0 0 0 0 0 0 0 0 1 0
126
+
127
+ 36 0 0 0 0 0 0 0 0 0 1
128
+
129
+ 37 0 0 0 0 0 0 0 0 0 1
130
+
131
+ 38 0 0 0 0 0 0 0 0 0 1
132
+
133
+ 39 0 0 0 0 0 0 0 0 0 1
134
+
135
+
136
+
137
+ model = BinaryRelevance(RandomForestClassifier(random_state = 0, n_estimators = 500))
138
+
139
+ model.fit(df1_x, df1_y)
140
+
141
+
142
+
143
+ df2_x = df2.copy()
144
+
145
+ df2_x = df2_x.loc[:, "x1":"x777"]
146
+
147
+ df2_x = df2_x.where(df2_x>0, 0)
148
+
149
+
150
+
151
+ print(model.predict_proba(df2_x))
152
+
153
+
154
+
155
+ >>> print(model.predict_proba(df2_x))
156
+
157
+ (0, 0) 0.768
158
+
159
+ (0, 1) 0.098
160
+
161
+ (0, 2) 0.032
162
+
163
+ (0, 3) 0.036
164
+
165
+ (0, 4) 0.048
166
+
167
+ (0, 7) 0.084
168
+
169
+ (0, 8) 0.056
170
+
171
+ (0, 9) 0.012
172
+
173
+ (1, 0) 0.04
174
+
175
+ (1, 1) 0.042
176
+
177
+ (1, 2) 0.732
178
+
179
+ (1, 3) 0.242
180
+
181
+ (1, 4) 0.024
182
+
183
+ (1, 5) 0.008
184
+
185
+ (1, 7) 0.038
186
+
187
+ (1, 8) 0.1
188
+
189
+ (1, 9) 0.014
190
+
191
+ (2, 0) 0.012
192
+
193
+ (2, 1) 0.024
194
+
195
+ (2, 2) 0.072
196
+
197
+ (2, 3) 0.012
198
+
199
+ (2, 4) 0.818
200
+
201
+ (2, 5) 0.09
202
+
203
+ (2, 7) 0.036
204
+
205
+ (2, 8) 0.028
206
+
207
+ (3, 0) 0.002
208
+
209
+ (3, 1) 0.024
210
+
211
+ (3, 2) 0.026
212
+
213
+ (3, 3) 0.004
214
+
215
+ (3, 4) 0.002
216
+
217
+ (3, 5) 0.004
218
+
219
+ (3, 6) 0.002
220
+
221
+ (3, 7) 0.61
222
+
223
+ (3, 8) 0.03
224
+
225
+ (3, 9) 0.006
226
+
227
+ (4, 0) 0.008
228
+
229
+ (4, 1) 0.04
230
+
231
+ (4, 2) 0.118
232
+
233
+ (4, 3) 0.016
234
+
235
+ (4, 4) 0.034
236
+
237
+ (4, 7) 0.032
238
+
239
+ (4, 8) 0.518
240
+
241
+ (4, 9) 0.22
242
+
243
+ (5, 0) 0.102
244
+
245
+ (5, 1) 0.08
246
+
247
+ (5, 2) 0.252
248
+
249
+ (5, 3) 0.062
250
+
251
+ (5, 4) 0.034
252
+
253
+ (5, 7) 0.082
254
+
255
+ (5, 8) 0.036
256
+
257
+ (5, 9) 0.02
258
+
259
+ (6, 0) 0.018
260
+
261
+ (6, 1) 0.032
262
+
263
+ (6, 2) 0.316
264
+
265
+ (6, 3) 0.102
266
+
267
+ (6, 4) 0.062
268
+
269
+ (6, 7) 0.052
270
+
271
+ (6, 8) 0.11
272
+
273
+ (6, 9) 0.016
274
+
275
+ (7, 0) 0.002
276
+
277
+ (7, 1) 0.026
278
+
279
+ (7, 2) 0.002
280
+
281
+ (7, 3) 0.032
282
+
283
+ (7, 4) 0.002
284
+
285
+ (7, 5) 0.084
286
+
287
+ (7, 6) 0.038
288
+
289
+ (7, 7) 0.322
290
+
291
+ (7, 8) 0.024
292
+
293
+ (7, 9) 0.002
294
+
295
+ (8, 0) 0.004
296
+
297
+ (8, 1) 0.046
298
+
299
+ (8, 2) 0.068
300
+
301
+ (8, 3) 0.028
302
+
303
+ (8, 4) 0.004
304
+
305
+ (8, 5) 0.008
306
+
307
+ (8, 7) 0.078
308
+
309
+ (8, 8) 0.158
310
+
311
+ (8, 9) 0.072
312
+
313
+ (9, 0) 0.13
314
+
315
+ (9, 1) 0.028
316
+
317
+ (9, 2) 0.046
318
+
319
+ (9, 3) 0.028
320
+
321
+ (9, 4) 0.004
322
+
323
+ (9, 6) 0.004
324
+
325
+ (9, 7) 0.088
326
+
327
+ (9, 8) 0.136
328
+
329
+ (9, 9) 0.038
330
+
331
+ (10, 0) 0.02
332
+
333
+ (10, 1) 0.092
334
+
335
+ (10, 2) 0.314
336
+
337
+ (10, 3) 0.034
338
+
339
+ (10, 4) 0.018
340
+
341
+ (10, 7) 0.072
342
+
343
+ (10, 8) 0.06
344
+
345
+ (10, 9) 0.036
346
+
347
+ (11, 0) 0.016
348
+
349
+ (11, 1) 0.006
350
+
351
+ (11, 2) 0.174
352
+
353
+ (11, 3) 0.054
354
+
355
+ (11, 4) 0.056
356
+
357
+ (11, 5) 0.028
358
+
359
+ (11, 6) 0.002
360
+
361
+ (11, 7) 0.052
362
+
363
+ (11, 8) 0.156
364
+
365
+ (11, 9) 0.012
366
+
367
+ (12, 0) 0.122
368
+
369
+ (12, 1) 0.064
370
+
371
+ (12, 2) 0.232
372
+
373
+ (12, 3) 0.094
374
+
375
+ (12, 4) 0.036
376
+
377
+ (12, 7) 0.088
378
+
379
+ (12, 8) 0.05
380
+
381
+ (12, 9) 0.038
382
+
383
+ (13, 0) 0.016
384
+
385
+ (13, 1) 0.006
386
+
387
+ (13, 2) 0.256
388
+
389
+ (13, 3) 0.08
390
+
391
+ (13, 4) 0.042
392
+
393
+ (13, 7) 0.054
394
+
395
+ (13, 8) 0.132
396
+
397
+ (13, 9) 0.03
398
+
399
+ (14, 0) 0.002
400
+
401
+ (14, 1) 0.008
402
+
403
+ (14, 2) 0.016
404
+
405
+ (14, 3) 0.012
406
+
407
+ (14, 4) 0.038
408
+
409
+ (14, 5) 0.002
410
+
411
+ (14, 6) 0.004
412
+
413
+ (14, 7) 0.158
414
+
415
+ (14, 8) 0.164
416
+
417
+ (14, 9) 0.006
418
+
419
+ (15, 0) 0.012
420
+
421
+ (15, 1) 0.022
422
+
423
+ (15, 2) 0.104
424
+
425
+ (15, 3) 0.014
426
+
427
+ (15, 4) 0.03
428
+
429
+ (15, 5) 0.002
430
+
431
+ (15, 6) 0.012
432
+
433
+ (15, 7) 0.1
434
+
435
+ (15, 8) 0.186
436
+
437
+ (15, 9) 0.074
438
+
439
+ (16, 0) 0.062
440
+
441
+ (16, 1) 0.034
442
+
443
+ (16, 2) 0.1
444
+
445
+ (16, 3) 0.094
446
+
447
+ (16, 4) 0.04
448
+
449
+ (16, 7) 0.112
450
+
451
+ (16, 8) 0.042
452
+
453
+ (16, 9) 0.03
454
+
455
+ (17, 0) 0.022
456
+
457
+ (17, 1) 0.012
458
+
459
+ (17, 2) 0.194
460
+
461
+ (17, 3) 0.05
462
+
463
+ (17, 4) 0.072
464
+
465
+ (17, 5) 0.006
466
+
467
+ (17, 6) 0.008
468
+
469
+ (17, 7) 0.05
470
+
471
+ (17, 8) 0.03
472
+
473
+ (17, 9) 0.022
474
+
475
+ (18, 0) 0.002
476
+
477
+ (18, 1) 0.034
478
+
479
+ (18, 2) 0.028
480
+
481
+ (18, 3) 0.016
482
+
483
+ (18, 4) 0.008
484
+
485
+ (18, 5) 0.002
486
+
487
+ (18, 6) 0.002
488
+
489
+ (18, 7) 0.08
490
+
491
+ (18, 8) 0.096
492
+
493
+ (18, 9) 0.02
494
+
495
+ (19, 0) 0.024
496
+
497
+ (19, 1) 0.042
498
+
499
+ (19, 2) 0.19
500
+
501
+ (19, 3) 0.064
502
+
503
+ (19, 4) 0.02
504
+
505
+ (19, 5) 0.002
506
+
507
+ (19, 7) 0.118
508
+
509
+ (19, 8) 0.104
510
+
511
+ (19, 9) 0.048
512
+
513
+
514
+
515
+ 例えば(0, 5)や(0, 6)の確率値が算出されていません。
16
516
 
17
517
 
18
518