forked from shogun-toolbox/shogun
-
Notifications
You must be signed in to change notification settings - Fork 1
/
GUIClassifier.cpp
1625 lines (1403 loc) · 39.9 KB
/
GUIClassifier.cpp
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
/*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 3 of the License, or
* (at your option) any later version.
*
* Written (W) 1999-2009 Soeren Sonnenburg
* Written (W) 1999-2008 Gunnar Raetsch
* Copyright (C) 1999-2009 Fraunhofer Institute FIRST and Max-Planck-Society
*/
#include <shogun/ui/GUIClassifier.h>
#include <shogun/ui/SGInterface.h>
#include <shogun/lib/config.h>
#include <shogun/io/SGIO.h>
#include <shogun/features/SparseFeatures.h>
#include <shogun/features/RealFileFeatures.h>
#include <shogun/features/Labels.h>
#include <shogun/kernel/AUCKernel.h>
#include <shogun/classifier/KNN.h>
#include <shogun/clustering/KMeans.h>
#include <shogun/clustering/Hierarchical.h>
#include <shogun/classifier/PluginEstimate.h>
#include <shogun/classifier/LDA.h>
#include <shogun/classifier/LPM.h>
#include <shogun/classifier/LPBoost.h>
#include <shogun/classifier/Perceptron.h>
#include <shogun/machine/LinearMachine.h>
#ifdef USE_SVMLIGHT
#include <shogun/classifier/svm/SVMLight.h>
#include <shogun/classifier/svm/SVMLightOneClass.h>
#include <shogun/regression/svr/SVRLight.h>
#endif //USE_SVMLIGHT
#include <shogun/classifier/mkl/MKLClassification.h>
#include <shogun/regression/svr/MKLRegression.h>
#include <shogun/classifier/mkl/MKLOneClass.h>
#include <shogun/classifier/mkl/MKLMultiClass.h>
#include <shogun/classifier/svm/LibSVM.h>
#include <shogun/classifier/svm/LaRank.h>
#include <shogun/classifier/svm/GPBTSVM.h>
#include <shogun/classifier/svm/LibSVMOneClass.h>
#include <shogun/classifier/svm/LibSVMMultiClass.h>
#include <shogun/regression/svr/LibSVR.h>
#include <shogun/regression/KernelRidgeRegression.h>
#include <shogun/classifier/svm/LibLinear.h>
#include <shogun/classifier/svm/MPDSVM.h>
#include <shogun/classifier/svm/GNPPSVM.h>
#include <shogun/classifier/svm/GMNPSVM.h>
#include <shogun/classifier/svm/ScatterSVM.h>
#include <shogun/classifier/svm/SVMLin.h>
#include <shogun/classifier/svm/SubGradientSVM.h>
#include <shogun/classifier/SubGradientLPM.h>
#include <shogun/classifier/svm/SVMOcas.h>
#include <shogun/classifier/svm/SVMSGD.h>
#include <shogun/classifier/svm/WDSVMOcas.h>
using namespace shogun;
CGUIClassifier::CGUIClassifier(CSGInterface* ui_)
: CSGObject(), ui(ui_)
{
constraint_generator=NULL;
classifier=NULL;
max_train_time=0;
// Perceptron parameters
perceptron_learnrate=0.1;
perceptron_maxiter=1000;
// SVM parameters
svm_qpsize=41;
svm_bufsize=3000;
svm_max_qpsize=1000;
mkl_norm=1;
ent_lambda=0;
mkl_block_norm=4;
svm_C1=1;
svm_C2=1;
C_mkl=0;
mkl_use_interleaved=true;
svm_weight_epsilon=1e-5;
svm_epsilon=1e-5;
svm_tube_epsilon=1e-2;
svm_nu=0.5;
svm_use_shrinking = true ;
svm_use_bias = true;
svm_use_batch_computation = true ;
svm_use_linadd = true ;
svm_do_auc_maximization = false ;
// KRR parameters
krr_tau=1;
solver_type=ST_AUTO;
}
CGUIClassifier::~CGUIClassifier()
{
SG_UNREF(classifier);
SG_UNREF(constraint_generator);
}
bool CGUIClassifier::new_classifier(char* name, int32_t d, int32_t from_d)
{
if (strcmp(name,"LIBSVM_ONECLASS")==0)
{
SG_UNREF(classifier);
classifier = new CLibSVMOneClass();
SG_INFO("created SVMlibsvm object for oneclass\n");
}
else if (strcmp(name,"LIBSVM_MULTICLASS")==0)
{
SG_UNREF(classifier);
classifier = new CLibSVMMultiClass();
SG_INFO("created SVMlibsvm object for multiclass\n");
}
else if (strcmp(name,"LIBSVM_NUMULTICLASS")==0)
{
SG_UNREF(classifier);
classifier= new CLibSVMMultiClass(LIBSVM_NU_SVC);
SG_INFO("created SVMlibsvm object for multiclass\n") ;
}
#ifdef USE_SVMLIGHT
else if (strcmp(name,"SCATTERSVM_NO_BIAS_SVMLIGHT")==0)
{
SG_UNREF(classifier);
classifier= new CScatterSVM(NO_BIAS_SVMLIGHT);
SG_INFO("created ScatterSVM NO BIAS SVMLIGHT object\n") ;
}
#endif //USE_SVMLIGHT
else if (strcmp(name,"SCATTERSVM_NO_BIAS_LIBSVM")==0)
{
SG_UNREF(classifier);
classifier= new CScatterSVM(NO_BIAS_LIBSVM);
SG_INFO("created ScatterSVM NO BIAS LIBSVM object\n") ;
}
else if (strcmp(name,"SCATTERSVM_TESTRULE1")==0)
{
SG_UNREF(classifier);
classifier= new CScatterSVM(TEST_RULE1);
SG_INFO("created ScatterSVM TESTRULE1 object\n") ;
}
else if (strcmp(name,"SCATTERSVM_TESTRULE2")==0)
{
SG_UNREF(classifier);
classifier= new CScatterSVM(TEST_RULE2);
SG_INFO("created ScatterSVM TESTRULE2 object\n") ;
}
else if (strcmp(name,"LIBSVM_NU")==0)
{
SG_UNREF(classifier);
classifier= new CLibSVM(LIBSVM_NU_SVC);
SG_INFO("created SVMlibsvm object\n") ;
}
else if (strcmp(name,"LIBSVM")==0)
{
SG_UNREF(classifier);
classifier= new CLibSVM();
SG_INFO("created SVMlibsvm object\n") ;
}
else if (strcmp(name,"LARANK")==0)
{
SG_UNREF(classifier);
classifier= new CLaRank();
SG_INFO("created LaRank object\n") ;
}
#ifdef USE_SVMLIGHT
else if ((strcmp(name,"LIGHT")==0) || (strcmp(name,"SVMLIGHT")==0))
{
SG_UNREF(classifier);
classifier= new CSVMLight();
SG_INFO("created SVMLight object\n") ;
}
else if (strcmp(name,"SVMLIGHT_ONECLASS")==0)
{
SG_UNREF(classifier);
classifier= new CSVMLightOneClass();
SG_INFO("created SVMLightOneClass object\n") ;
}
else if (strcmp(name,"SVRLIGHT")==0)
{
SG_UNREF(classifier);
classifier= new CSVRLight();
SG_INFO("created SVRLight object\n") ;
}
#endif //USE_SVMLIGHT
else if (strcmp(name,"GPBTSVM")==0)
{
SG_UNREF(classifier);
classifier= new CGPBTSVM();
SG_INFO("created GPBT-SVM object\n") ;
}
else if (strcmp(name,"MPDSVM")==0)
{
SG_UNREF(classifier);
classifier= new CMPDSVM();
SG_INFO("created MPD-SVM object\n") ;
}
else if (strcmp(name,"GNPPSVM")==0)
{
SG_UNREF(classifier);
classifier= new CGNPPSVM();
SG_INFO("created GNPP-SVM object\n") ;
}
else if (strcmp(name,"GMNPSVM")==0)
{
SG_UNREF(classifier);
classifier= new CGMNPSVM();
SG_INFO("created GMNP-SVM object\n") ;
}
else if (strcmp(name,"LIBSVR")==0)
{
SG_UNREF(classifier);
classifier= new CLibSVR();
SG_INFO("created SVRlibsvm object\n") ;
}
#ifdef HAVE_LAPACK
else if (strcmp(name, "KRR")==0)
{
SG_UNREF(classifier);
classifier=new CKernelRidgeRegression(krr_tau, ui->ui_kernel->get_kernel(),
ui->ui_labels->get_train_labels());
SG_INFO("created KRR object %p\n", classifier);
}
#endif //HAVE_LAPACK
else if (strcmp(name,"PERCEPTRON")==0)
{
SG_UNREF(classifier);
classifier= new CPerceptron();
SG_INFO("created Perceptron object\n") ;
}
#ifdef HAVE_LAPACK
else if (strncmp(name,"LIBLINEAR",9)==0)
{
LIBLINEAR_SOLVER_TYPE st=L2R_LR;
if (strcmp(name,"LIBLINEAR_L2R_LR")==0)
{
st=L2R_LR;
SG_INFO("created LibLinear l2 regularized logistic regression object\n") ;
}
else if (strcmp(name,"LIBLINEAR_L2R_L2LOSS_SVC_DUAL")==0)
{
st=L2R_L2LOSS_SVC_DUAL;
SG_INFO("created LibLinear l2 regularized l2 loss SVM dual object\n") ;
}
else if (strcmp(name,"LIBLINEAR_L2R_L2LOSS_SVC")==0)
{
st=L2R_L2LOSS_SVC;
SG_INFO("created LibLinear l2 regularized l2 loss SVM primal object\n") ;
}
else if (strcmp(name,"LIBLINEAR_L1R_L2LOSS_SVC")==0)
{
st=L1R_L2LOSS_SVC;
SG_INFO("created LibLinear l1 regularized l2 loss SVM primal object\n") ;
}
else if (strcmp(name,"LIBLINEAR_L2R_L1LOSS_SVC_DUAL")==0)
{
st=L2R_L1LOSS_SVC_DUAL;
SG_INFO("created LibLinear l2 regularized l1 loss dual SVM object\n") ;
}
else
SG_ERROR("unknown liblinear type\n");
SG_UNREF(classifier);
classifier= new CLibLinear(st);
((CLibLinear*) classifier)->set_C(svm_C1, svm_C2);
((CLibLinear*) classifier)->set_epsilon(svm_epsilon);
((CLibLinear*) classifier)->set_bias_enabled(svm_use_bias);
}
else if (strcmp(name,"LDA")==0)
{
SG_UNREF(classifier);
classifier= new CLDA();
SG_INFO("created LDA object\n") ;
}
#endif //HAVE_LAPACK
#ifdef USE_CPLEX
else if (strcmp(name,"LPM")==0)
{
SG_UNREF(classifier);
classifier= new CLPM();
((CLPM*) classifier)->set_C(svm_C1, svm_C2);
((CLPM*) classifier)->set_epsilon(svm_epsilon);
((CLPM*) classifier)->set_bias_enabled(svm_use_bias);
((CLPM*) classifier)->set_max_train_time(max_train_time);
SG_INFO("created LPM object\n") ;
}
else if (strcmp(name,"LPBOOST")==0)
{
SG_UNREF(classifier);
classifier= new CLPBoost();
((CLPBoost*) classifier)->set_C(svm_C1, svm_C2);
((CLPBoost*) classifier)->set_epsilon(svm_epsilon);
((CLPBoost*) classifier)->set_bias_enabled(svm_use_bias);
((CLPBoost*) classifier)->set_max_train_time(max_train_time);
SG_INFO("created LPBoost object\n") ;
}
else if (strcmp(name,"SUBGRADIENTLPM")==0)
{
SG_UNREF(classifier);
classifier= new CSubGradientLPM();
((CSubGradientLPM*) classifier)->set_bias_enabled(svm_use_bias);
((CSubGradientLPM*) classifier)->set_qpsize(svm_qpsize);
((CSubGradientLPM*) classifier)->set_qpsize_max(svm_max_qpsize);
((CSubGradientLPM*) classifier)->set_C(svm_C1, svm_C2);
((CSubGradientLPM*) classifier)->set_epsilon(svm_epsilon);
((CSubGradientLPM*) classifier)->set_max_train_time(max_train_time);
SG_INFO("created Subgradient LPM object\n") ;
}
#endif //USE_CPLEX
else if (strncmp(name,"KNN", strlen("KNN"))==0)
{
SG_UNREF(classifier);
classifier= new CKNN();
SG_INFO("created KNN object\n") ;
}
else if (strncmp(name,"KMEANS", strlen("KMEANS"))==0)
{
SG_UNREF(classifier);
classifier= new CKMeans();
SG_INFO("created KMeans object\n") ;
}
else if (strncmp(name,"HIERARCHICAL", strlen("HIERARCHICAL"))==0)
{
SG_UNREF(classifier);
classifier= new CHierarchical();
SG_INFO("created Hierarchical clustering object\n") ;
}
else if (strcmp(name,"SVMLIN")==0)
{
SG_UNREF(classifier);
classifier= new CSVMLin();
((CSVMLin*) classifier)->set_C(svm_C1, svm_C2);
((CSVMLin*) classifier)->set_epsilon(svm_epsilon);
((CSVMLin*) classifier)->set_bias_enabled(svm_use_bias);
SG_INFO("created SVMLin object\n") ;
}
else if (strcmp(name,"SUBGRADIENTSVM")==0)
{
SG_UNREF(classifier);
classifier= new CSubGradientSVM();
((CSubGradientSVM*) classifier)->set_bias_enabled(svm_use_bias);
((CSubGradientSVM*) classifier)->set_qpsize(svm_qpsize);
((CSubGradientSVM*) classifier)->set_qpsize_max(svm_max_qpsize);
((CSubGradientSVM*) classifier)->set_C(svm_C1, svm_C2);
((CSubGradientSVM*) classifier)->set_epsilon(svm_epsilon);
((CSubGradientSVM*) classifier)->set_max_train_time(max_train_time);
SG_INFO("created Subgradient SVM object\n") ;
}
else if (strncmp(name,"WDSVMOCAS", strlen("WDSVMOCAS"))==0)
{
SG_UNREF(classifier);
classifier= new CWDSVMOcas(SVM_OCAS);
((CWDSVMOcas*) classifier)->set_bias_enabled(svm_use_bias);
((CWDSVMOcas*) classifier)->set_degree(d, from_d);
((CWDSVMOcas*) classifier)->set_C(svm_C1, svm_C2);
((CWDSVMOcas*) classifier)->set_epsilon(svm_epsilon);
((CWDSVMOcas*) classifier)->set_bufsize(svm_bufsize);
SG_INFO("created Weighted Degree Kernel SVM Ocas(OCAS) object of order %d (from order:%d)\n", d, from_d) ;
}
else if (strcmp(name,"SVMOCAS")==0)
{
SG_UNREF(classifier);
classifier= new CSVMOcas(SVM_OCAS);
((CSVMOcas*) classifier)->set_C(svm_C1, svm_C2);
((CSVMOcas*) classifier)->set_epsilon(svm_epsilon);
((CSVMOcas*) classifier)->set_bufsize(svm_bufsize);
((CSVMOcas*) classifier)->set_bias_enabled(svm_use_bias);
SG_INFO("created SVM Ocas(OCAS) object\n") ;
}
else if (strcmp(name,"SVMSGD")==0)
{
SG_UNREF(classifier);
classifier= new CSVMSGD(svm_C1);
((CSVMSGD*) classifier)->set_bias_enabled(svm_use_bias);
SG_INFO("created SVM SGD object\n") ;
}
else if (strcmp(name,"SVMBMRM")==0 || (strcmp(name,"SVMPERF")==0))
{
SG_UNREF(classifier);
classifier= new CSVMOcas(SVM_BMRM);
((CSVMOcas*) classifier)->set_C(svm_C1, svm_C2);
((CSVMOcas*) classifier)->set_epsilon(svm_epsilon);
((CSVMOcas*) classifier)->set_bufsize(svm_bufsize);
((CSVMOcas*) classifier)->set_bias_enabled(svm_use_bias);
SG_INFO("created SVM Ocas(BMRM/PERF) object\n") ;
}
else if (strcmp(name,"MKL_CLASSIFICATION")==0)
{
SG_UNREF(classifier);
classifier= new CMKLClassification();
}
//else if (strcmp(name,"MKL_MULTICLASS")==0)
//{
// SG_UNREF(classifier);
// classifier= new CMKLClassification();
//}
else if (strcmp(name,"MKL_ONECLASS")==0)
{
SG_UNREF(classifier);
classifier= new CMKLOneClass();
}
else if (strcmp(name,"MKL_MULTICLASS")==0)
{
SG_UNREF(classifier);
classifier= new CMKLMultiClass();
}
else if (strcmp(name,"MKL_REGRESSION")==0)
{
SG_UNREF(classifier);
classifier= new CMKLRegression();
}
else
{
SG_ERROR("Unknown classifier %s.\n", name);
return false;
}
SG_REF(classifier);
return (classifier!=NULL);
}
bool CGUIClassifier::train_mkl_multiclass()
{
CMKLMultiClass* mkl= (CMKLMultiClass*) classifier;
if (!mkl)
SG_ERROR("No MKL available.\n");
CLabels* trainlabels=ui->ui_labels->get_train_labels();
if (!trainlabels)
SG_ERROR("No trainlabels available.\n");
CKernel* kernel=ui->ui_kernel->get_kernel();
if (!kernel)
SG_ERROR("No kernel available.\n");
bool success=ui->ui_kernel->init_kernel("TRAIN");
if (!success || !ui->ui_kernel->is_initialized() || !kernel->has_features())
SG_ERROR("Kernel not initialized / no train features available.\n");
int32_t num_vec=kernel->get_num_vec_lhs();
if (trainlabels->get_num_labels() != num_vec)
SG_ERROR("Number of train labels (%d) and training vectors (%d) differs!\n", trainlabels->get_num_labels(), num_vec);
SG_INFO("Starting MC-MKL training on %ld vectors using C1=%lf C2=%lf epsilon=%lf\n", num_vec, svm_C1, svm_C2, svm_epsilon);
mkl->set_mkl_epsilon(svm_weight_epsilon);
mkl->set_mkl_norm(mkl_norm);
//mkl->set_max_num_mkliters(-1);
mkl->set_solver_type(solver_type);
mkl->set_bias_enabled(svm_use_bias);
mkl->set_epsilon(svm_epsilon);
mkl->set_max_train_time(max_train_time);
mkl->set_tube_epsilon(svm_tube_epsilon);
mkl->set_nu(svm_nu);
mkl->set_C(svm_C1, svm_C2);
mkl->set_qpsize(svm_qpsize);
mkl->set_shrinking_enabled(svm_use_shrinking);
mkl->set_linadd_enabled(svm_use_linadd);
mkl->set_batch_computation_enabled(svm_use_batch_computation);
((CKernelMachine*) mkl)->set_labels(trainlabels);
((CKernelMachine*) mkl)->set_kernel(kernel);
return mkl->train();
}
bool CGUIClassifier::train_mkl()
{
CMKL* mkl= (CMKL*) classifier;
if (!mkl)
SG_ERROR("No SVM available.\n");
bool oneclass=(mkl->get_classifier_type()==CT_LIBSVMONECLASS);
CLabels* trainlabels=NULL;
if(!oneclass)
trainlabels=ui->ui_labels->get_train_labels();
else
SG_INFO("Training one class mkl.\n");
if (!trainlabels && !oneclass)
SG_ERROR("No trainlabels available.\n");
CKernel* kernel=ui->ui_kernel->get_kernel();
if (!kernel)
SG_ERROR("No kernel available.\n");
bool success=ui->ui_kernel->init_kernel("TRAIN");
if (!success || !ui->ui_kernel->is_initialized() || !kernel->has_features())
SG_ERROR("Kernel not initialized.\n");
int32_t num_vec=kernel->get_num_vec_lhs();
if (!oneclass && trainlabels->get_num_labels() != num_vec)
SG_ERROR("Number of train labels (%d) and training vectors (%d) differs!\n", trainlabels->get_num_labels(), num_vec);
SG_INFO("Starting SVM training on %ld vectors using C1=%lf C2=%lf epsilon=%lf\n", num_vec, svm_C1, svm_C2, svm_epsilon);
if (constraint_generator)
mkl->set_constraint_generator(constraint_generator);
mkl->set_solver_type(solver_type);
mkl->set_bias_enabled(svm_use_bias);
mkl->set_epsilon(svm_epsilon);
mkl->set_max_train_time(max_train_time);
mkl->set_tube_epsilon(svm_tube_epsilon);
mkl->set_nu(svm_nu);
mkl->set_C(svm_C1, svm_C2);
mkl->set_qpsize(svm_qpsize);
mkl->set_shrinking_enabled(svm_use_shrinking);
mkl->set_linadd_enabled(svm_use_linadd);
mkl->set_batch_computation_enabled(svm_use_batch_computation);
mkl->set_mkl_epsilon(svm_weight_epsilon);
mkl->set_mkl_norm(mkl_norm);
mkl->set_elasticnet_lambda(ent_lambda);
mkl->set_mkl_block_norm(mkl_block_norm);
mkl->set_C_mkl(C_mkl);
mkl->set_interleaved_optimization_enabled(mkl_use_interleaved);
if (svm_do_auc_maximization)
{
CAUCKernel* auc_kernel = new CAUCKernel(10, kernel);
CLabels* auc_labels= auc_kernel->setup_auc_maximization(trainlabels);
((CKernelMachine*) mkl)->set_labels(auc_labels);
((CKernelMachine*) mkl)->set_kernel(auc_kernel);
SG_UNREF(auc_labels);
}
else
{
if(!oneclass)
((CKernelMachine*) mkl)->set_labels(trainlabels);
((CKernelMachine*) mkl)->set_kernel(kernel);
}
bool result=mkl->train();
return result;
}
bool CGUIClassifier::train_svm()
{
CSVM* svm= (CSVM*) classifier;
if (!svm)
SG_ERROR("No SVM available.\n");
bool oneclass=(svm->get_classifier_type()==CT_LIBSVMONECLASS);
CLabels* trainlabels=NULL;
if(!oneclass)
trainlabels=ui->ui_labels->get_train_labels();
else
SG_INFO("Training one class svm.\n");
if (!trainlabels && !oneclass)
SG_ERROR("No trainlabels available.\n");
CKernel* kernel=ui->ui_kernel->get_kernel();
if (!kernel)
SG_ERROR("No kernel available.\n");
bool success=ui->ui_kernel->init_kernel("TRAIN");
if (!success || !ui->ui_kernel->is_initialized() || !kernel->has_features())
SG_ERROR("Kernel not initialized / no train features available.\n");
int32_t num_vec=kernel->get_num_vec_lhs();
if (!oneclass && trainlabels->get_num_labels() != num_vec)
SG_ERROR("Number of train labels (%d) and training vectors (%d) differs!\n", trainlabels->get_num_labels(), num_vec);
SG_INFO("Starting SVM training on %ld vectors using C1=%lf C2=%lf epsilon=%lf\n", num_vec, svm_C1, svm_C2, svm_epsilon);
svm->set_solver_type(solver_type);
svm->set_bias_enabled(svm_use_bias);
svm->set_epsilon(svm_epsilon);
svm->set_max_train_time(max_train_time);
svm->set_tube_epsilon(svm_tube_epsilon);
svm->set_nu(svm_nu);
svm->set_C(svm_C1, svm_C2);
svm->set_qpsize(svm_qpsize);
svm->set_shrinking_enabled(svm_use_shrinking);
svm->set_linadd_enabled(svm_use_linadd);
svm->set_batch_computation_enabled(svm_use_batch_computation);
if(svm->get_classifier_type()==CT_MKLMULTICLASS)
{
((CMKLMultiClass *)svm)->set_mkl_epsilon(svm_weight_epsilon );
}
if (svm_do_auc_maximization)
{
CAUCKernel* auc_kernel = new CAUCKernel(10, kernel);
CLabels* auc_labels= auc_kernel->setup_auc_maximization(trainlabels);
((CKernelMachine*) svm)->set_labels(auc_labels);
((CKernelMachine*) svm)->set_kernel(auc_kernel);
SG_UNREF(auc_labels);
}
else
{
if(!oneclass)
((CKernelMachine*) svm)->set_labels(trainlabels);
((CKernelMachine*) svm)->set_kernel(kernel);
}
bool result=svm->train();
return result;
}
bool CGUIClassifier::train_clustering(int32_t k, int32_t max_iter)
{
bool result=false;
CDistance* distance=ui->ui_distance->get_distance();
if (!distance)
SG_ERROR("No distance available\n");
if (!ui->ui_distance->init_distance("TRAIN"))
SG_ERROR("Initializing distance with train features failed.\n");
((CDistanceMachine*) classifier)->set_distance(distance);
EClassifierType type=classifier->get_classifier_type();
switch (type)
{
case CT_KMEANS:
{
((CKMeans*) classifier)->set_k(k);
((CKMeans*) classifier)->set_max_iter(max_iter);
result=((CKMeans*) classifier)->train();
break;
}
case CT_HIERARCHICAL:
{
((CHierarchical*) classifier)->set_merges(k);
result=((CHierarchical*) classifier)->train();
break;
}
default:
SG_ERROR("Unknown clustering type %d\n", type);
}
return result;
}
bool CGUIClassifier::train_knn(int32_t k)
{
CLabels* trainlabels=ui->ui_labels->get_train_labels();
CDistance* distance=ui->ui_distance->get_distance();
bool result=false;
if (trainlabels)
{
if (distance)
{
if (!ui->ui_distance->init_distance("TRAIN"))
SG_ERROR("Initializing distance with train features failed.\n");
((CKNN*) classifier)->set_labels(trainlabels);
((CKNN*) classifier)->set_distance(distance);
((CKNN*) classifier)->set_k(k);
result=((CKNN*) classifier)->train();
}
else
SG_ERROR("No distance available.\n");
}
else
SG_ERROR("No labels available\n");
return result;
}
bool CGUIClassifier::train_krr()
{
#ifdef HAVE_LAPACK
CKernelRidgeRegression* krr= (CKernelRidgeRegression*) classifier;
if (!krr)
SG_ERROR("No SVM available.\n");
CLabels* trainlabels=NULL;
trainlabels=ui->ui_labels->get_train_labels();
if (!trainlabels)
SG_ERROR("No trainlabels available.\n");
CKernel* kernel=ui->ui_kernel->get_kernel();
if (!kernel)
SG_ERROR("No kernel available.\n");
bool success=ui->ui_kernel->init_kernel("TRAIN");
if (!success || !ui->ui_kernel->is_initialized() || !kernel->has_features())
SG_ERROR("Kernel not initialized / no train features available.\n");
int32_t num_vec=kernel->get_num_vec_lhs();
if (trainlabels->get_num_labels() != num_vec)
SG_ERROR("Number of train labels (%d) and training vectors (%d) differs!\n", trainlabels->get_num_labels(), num_vec);
// Set training labels and kernel
krr->set_labels(trainlabels);
krr->set_kernel(kernel);
bool result=krr->train();
return result;
#else
return false;
#endif
}
bool CGUIClassifier::train_linear(float64_t gamma)
{
ASSERT(classifier);
EClassifierType ctype = classifier->get_classifier_type();
CFeatures* trainfeatures=ui->ui_features->get_train_features();
CLabels* trainlabels=ui->ui_labels->get_train_labels();
bool result=false;
if (!trainfeatures)
SG_ERROR("No trainfeatures available.\n");
if (!trainfeatures->has_property(FP_DOT))
SG_ERROR("Trainfeatures not based on DotFeatures.\n");
if (!trainlabels)
SG_ERROR("No labels available\n");
if (ctype==CT_PERCEPTRON)
{
((CPerceptron*) classifier)->set_learn_rate(perceptron_learnrate);
((CPerceptron*) classifier)->set_max_iter(perceptron_maxiter);
}
#ifdef HAVE_LAPACK
if (ctype==CT_LDA)
{
if (trainfeatures->get_feature_type()!=F_DREAL ||
trainfeatures->get_feature_class()!=C_SIMPLE)
SG_ERROR("LDA requires train features of class SIMPLE type REAL.\n");
((CLDA*) classifier)->set_gamma(gamma);
}
#endif
if (ctype==CT_SVMOCAS)
((CSVMOcas*) classifier)->set_C(svm_C1, svm_C2);
#ifdef HAVE_LAPACK
else if (ctype==CT_LIBLINEAR)
((CLibLinear*) classifier)->set_C(svm_C1, svm_C2);
#endif
else if (ctype==CT_SVMLIN)
((CSVMLin*) classifier)->set_C(svm_C1, svm_C2);
else if (ctype==CT_SVMSGD)
((CSVMSGD*) classifier)->set_C(svm_C1, svm_C2);
else if (ctype==CT_SUBGRADIENTSVM)
((CSubGradientSVM*) classifier)->set_C(svm_C1, svm_C2);
else if (ctype==CT_LPM || ctype==CT_LPBOOST)
{
if (trainfeatures->get_feature_class()!=C_SPARSE ||
trainfeatures->get_feature_type()!=F_DREAL)
SG_ERROR("LPM and LPBOOST require trainfeatures of class SPARSE type REAL.\n");
}
((CLinearMachine*) classifier)->set_labels(trainlabels);
((CLinearMachine*) classifier)->set_features((CSimpleFeatures<float64_t>*) trainfeatures);
result=((CLinearMachine*) classifier)->train();
return result;
}
bool CGUIClassifier::train_wdocas()
{
CFeatures* trainfeatures=ui->ui_features->get_train_features();
CLabels* trainlabels=ui->ui_labels->get_train_labels();
bool result=false;
if (!trainfeatures)
SG_ERROR("No trainfeatures available.\n");
if (trainfeatures->get_feature_class()!=C_STRING ||
trainfeatures->get_feature_type()!=F_BYTE )
SG_ERROR("Trainfeatures are not of class STRING type BYTE.\n");
if (!trainlabels)
SG_ERROR("No labels available.\n");
((CWDSVMOcas*) classifier)->set_labels(trainlabels);
((CWDSVMOcas*) classifier)->set_features((CStringFeatures<uint8_t>*) trainfeatures);
result=((CWDSVMOcas*) classifier)->train();
return result;
}
bool CGUIClassifier::load(char* filename, char* type)
{
bool result=false;
if (new_classifier(type))
{
FILE* model_file=fopen(filename, "r");
if (model_file)
{
if (classifier && classifier->load(model_file))
{
SG_DEBUG("file successfully read.\n");
result=true;
}
else
SG_ERROR("SVM/Classifier creation/loading failed on file %s.\n", filename);
fclose(model_file);
}
else
SG_ERROR("Opening file %s failed.\n", filename);
return result;
}
else
SG_ERROR("Type %s of SVM/Classifier unknown.\n", type);
return false;
}
bool CGUIClassifier::save(char* param)
{
bool result=false;
param=SGIO::skip_spaces(param);
if (classifier)
{
FILE* file=fopen(param, "w");
if ((!file) || (!classifier->save(file)))
printf("writing to file %s failed!\n", param);
else
{
printf("successfully written classifier into \"%s\" !\n", param);
result=true;
}
if (file)
fclose(file);
}
else
SG_ERROR("create classifier first\n");
return result;
}
bool CGUIClassifier::set_perceptron_parameters(
float64_t learnrate, int32_t maxiter)
{
if (learnrate<=0)
perceptron_learnrate=0.01;
else
perceptron_learnrate=learnrate;
if (maxiter<=0)
perceptron_maxiter=1000;
else
perceptron_maxiter=maxiter;
SG_INFO("Setting to perceptron parameters (learnrate %f and maxiter: %d\n", perceptron_learnrate, perceptron_maxiter);
return true;
}
bool CGUIClassifier::set_svm_epsilon(float64_t epsilon)
{
if (epsilon<0)
svm_epsilon=1e-4;
else
svm_epsilon=epsilon;
SG_INFO("Set to svm_epsilon=%f.\n", svm_epsilon);
return true;
}
bool CGUIClassifier::set_max_train_time(float64_t max)
{
if (max>0)
{
max_train_time=max;
SG_INFO("Set to max_train_time=%f.\n", max_train_time);
}
else
SG_INFO("Disabling max_train_time.\n");
return true;
}
bool CGUIClassifier::set_svr_tube_epsilon(float64_t tube_epsilon)
{
if (!classifier)
SG_ERROR("No regression method allocated\n");
if (classifier->get_classifier_type() != CT_LIBSVR &&
classifier->get_classifier_type() != CT_SVRLIGHT &&
classifier->get_classifier_type() != CT_MKLREGRESSION )
{
SG_ERROR("Underlying method not capable of SV-regression\n");
}
if (tube_epsilon<0)
svm_tube_epsilon=1e-2;
svm_tube_epsilon=tube_epsilon;
((CSVM*) classifier)->set_tube_epsilon(svm_tube_epsilon);
SG_INFO("Set to svr_tube_epsilon=%f.\n", svm_tube_epsilon);
return true;
}
bool CGUIClassifier::set_svm_nu(float64_t nu)
{
if (nu<0 || nu>1)
nu=0.5;
svm_nu=nu;
SG_INFO("Set to nu=%f.\n", svm_nu);
return true;
}
bool CGUIClassifier::set_svm_mkl_parameters(
float64_t weight_epsilon, float64_t C, float64_t norm)
{
if (weight_epsilon<0)
weight_epsilon=1e-4;
if (C<0)
C=0;
if (norm<0)
SG_ERROR("MKL norm >= 0\n");
svm_weight_epsilon=weight_epsilon;
C_mkl=C;
mkl_norm=norm;
SG_INFO("Set to weight_epsilon=%f.\n", svm_weight_epsilon);
SG_INFO("Set to C_mkl=%f.\n", C_mkl);
SG_INFO("Set to mkl_norm=%f.\n", mkl_norm);
return true;
}
bool CGUIClassifier::set_elasticnet_lambda(float64_t lambda)
{
if (lambda<0 || lambda>1)
SG_ERROR("0 <= ent_lambda <= 1\n");
ent_lambda = lambda;
return true;
}
bool CGUIClassifier::set_mkl_block_norm(float64_t mkl_bnorm)
{
if (mkl_bnorm<1)
SG_ERROR("1 <= mkl_block_norm <= inf\n");
mkl_block_norm=mkl_bnorm;
return true;
}
bool CGUIClassifier::set_svm_C(float64_t C1, float64_t C2)
{
if (C1<0)
svm_C1=1.0;
else
svm_C1=C1;
if (C2<0)
svm_C2=svm_C1;
else
svm_C2=C2;
SG_INFO("Set to C1=%f C2=%f.\n", svm_C1, svm_C2);
return true;
}
bool CGUIClassifier::set_svm_qpsize(int32_t qpsize)
{
if (qpsize<2)
svm_qpsize=41;
else
svm_qpsize=qpsize;
SG_INFO("Set qpsize to svm_qpsize=%d.\n", svm_qpsize);