forked from sherbold/replication-kit-tse-2017-benchmark
-
Notifications
You must be signed in to change notification settings - Fork 0
/
generate_results.R
849 lines (771 loc) · 46.1 KB
/
generate_results.R
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
# Script for the result evaluation and generation of plots.
#########################
# Definition of Globals #
#########################
# You may change these global variables for different purposes.
# You should at least check the MySQL configuration below
# MySQL Configuration -
MYSQL_HOST = "localhost"
MYSQL_PORT = 3306
MYSQL_DBNAME = "crosspare"
MYSQL_USER = "crosspare"
MYSQL_PASSWORT = "crosspare"
# Path for all generates plots
PLOT_PATH = "figures/"
# Path for all generated tables
TABLES_PATH = "tables/"
# If true the plots included in the article are generated
CREATEPLOTS = FALSE
# Defines wether Friedman-Nemenyi or ANOVA with Scott-Knott clustering is used for ranking
# TRUE = Friedman-Nemenyi, FALSE = Scott-Knott
NONPARAMETRIC = TRUE
# 1 for fine-grained Nemenyi ranking from the correction
# Any other number for the three-ranks approach from the original paper
NONPARAMETRICRANKINGMODE = 1
# If true CD charts for Friedman-Nemenyi tests are created
PRINTCDCHARTS = FALSE
CD_EXPORT_WIDTH = 10
#################################
# Install and load dependencies #
#################################
if (!require("RMySQL")) install.packages("RMySQL")
if (!require("ScottKnott")) install.packages("ScottKnott")
if (!require("ggplot2")) install.packages("ggplot2")
if (!require("gridExtra")) install.packages("gridExtra")
if (!require("xtable")) install.packages("xtable")
if (!require("reshape")) install.packages("reshape")
if (!require("gdata")) install.packages("gdata")
if (!require("stringr")) install.packages("stringr")
library(RMySQL)
library(ScottKnott)
library(ggplot2)
library(gridExtra)
library(xtable)
library(reshape)
library(gdata)
library(stringr)
# We have to increase the number of nested expressions, because the CD diagrams cannot be plotted otherwise
options(expressions=10000)
########################################################
# Definition of functions with result evaluation logic #
########################################################
evaluateCPDPBenchmark = function(metricNames, datasets) {
mydb = dbConnect(MySQL(), user=MYSQL_USER, password=MYSQL_PASSWORT, host=MYSQL_HOST, port=MYSQL_PORT, dbname=MYSQL_DBNAME)
dbListTables(mydb)
dbListFields(mydb, 'crosspare.results')
sqlStatement = "SELECT DISTINCT concat(substring(configurationName, 8), '-', classifier) as config FROM resultsView WHERE configurationName LIKE 'RELINK-%';"
rs = dbSendQuery(mydb, sqlStatement)
configurations <- fetch(rs, n=-1)
for( i in 1:length(datasets) ) {
dataset = datasets[i]
sqlStatement = paste("SELECT concat(substring(configurationName, ",
nchar(dataset)+2,
"), '-', classifier) as config, ",
paste(metricNames, collapse=", "),
" FROM resultsView WHERE configurationName LIKE '",
dataset, "-%';", sep = "")
rs = dbSendQuery(mydb, sqlStatement)
results = fetch(rs, n=-1)
results[is.na(results)] = 0
cnts = data.frame(config=unique(results$config), count=0)
for( j in 1:nrow(results) ) {
cnts[cnts$config==results$config[j],]$count = cnts[cnts$config==results$config[j],]$count+1
results$index[j] = cnts[cnts$config==results$config[j],]$count
}
if( "auc" %in% metricNames ) {
if( NONPARAMETRIC) {
# Friedman-Nemenyi Test
meanRanks = createMeanRankMat(results, "auc")
if( PRINTCDCHARTS ) {
printMetricName = "AUC"
plotObj = plotCDNemenyiFriedman(meanRanks, max(results$index), title=paste("Critical Distance Diagram for ", printMetricName, " and the ", dataset, " data", sep=""))
ggsave(filename = paste(PLOT_PATH,printMetricName,"_",dataset,".png",sep=""), plot=plotObj, device="png", width = CD_EXPORT_WIDTH, height = CD_EXPORT_WIDTH)
}
colnumber = ncol(configurations)+1
if( NONPARAMETRICRANKINGMODE==1 ) {
for( j in 1:nrow(configurations) ) {
if( configurations$config[j] %in% meanRanks$config) {
configurations[j,colnumber] = meanRanks[meanRanks$config==configurations$config[j],]$normRank
}
}
} else {
cdNemenyi = getNemenyiCD(0.05, length(unique(results$config)), max(results$index))
maxMeanRank = max(meanRanks$meanRank)
minMeanRank = min(meanRanks$meanRank)
# with CD within first results
if( maxMeanRank-minMeanRank<=cdNemenyi ) {
# all on first rank
configurations[, colnumber] = 1
}
else if(maxMeanRank-minMeanRank<=2*cdNemenyi) {
# only two ranks
firstRank = meanRanks$config[meanRanks$meanRank>=maxMeanRank-cdNemenyi]
secondRank = meanRanks$config[meanRanks$meanRank<maxMeanRank-cdNemenyi]
configurations[configurations$config %in% firstRank, colnumber] = 1
configurations[configurations$config %in% secondRank, colnumber] = 1-length(firstRank)/(nrow(meanRanks)-1)
} else {
# three ranks
firstRank <<- meanRanks$config[meanRanks$meanRank>=maxMeanRank-cdNemenyi]
secondRank <<- meanRanks$config[meanRanks$meanRank>minMeanRank+cdNemenyi & meanRanks$meanRank<maxMeanRank-cdNemenyi]
thirdRank <<- meanRanks$config[meanRanks$meanRank<=minMeanRank+cdNemenyi]
configurations[configurations$config %in% firstRank, colnumber] = 1
configurations[configurations$config %in% secondRank, colnumber] = 1-length(firstRank)/(nrow(meanRanks)-1)
configurations[configurations$config %in% thirdRank, colnumber] = 1-(length(firstRank)+length(secondRank))/(nrow(meanRanks)-1)
}
}
colnames(configurations)[colnumber] = paste("AUC",dataset,sep="_")
} else {
# AOV and Scott-Knott test
skresult = SK(aov( auc ~ config, results), sig.level=0.05)
colnumber = ncol(configurations)+1
higherRanked = 0
for( j in 1:length(unique(skresult$groups)) ) {
orderedNms = skresult$nms[skresult$ord]
configurations[configurations$config %in% orderedNms[skresult$groups==j], colnumber] = 1-(higherRanked/(nrow(configurations)-1))
higherRanked <- higherRanked+length(orderedNms[skresult$groups==j])
}
colnames(configurations)[colnumber] = paste("AUC",dataset,sep="_")
}
colnumber = ncol(configurations)+1
for( j in 1:nrow(configurations) ) {
configurations[j,colnumber] = mean(results[results$config==configurations$config[j],]$auc, na.rm = TRUE)
}
colnames(configurations)[colnumber] = paste("meanAUC",dataset,sep="_")
}
if( "fscore" %in% metricNames ) {
if (NONPARAMETRIC ) {
# Friedman-Nemenyi Test
meanRanks = createMeanRankMat(results, "fscore")
if( PRINTCDCHARTS ) {
printMetricName = "F-Measure"
plotObj = plotCDNemenyiFriedman(meanRanks, max(results$index), title=paste("Critical Distance Diagram for ", printMetricName, " and the ", dataset, " data", sep=""))
ggsave(filename = paste(PLOT_PATH,printMetricName,"_",dataset,".png",sep=""), plot=plotObj, device="png", width = CD_EXPORT_WIDTH, height = CD_EXPORT_WIDTH)
}
colnumber = ncol(configurations)+1
if( NONPARAMETRICRANKINGMODE==1 ) {
for( j in 1:nrow(configurations) ) {
if( configurations$config[j] %in% meanRanks$config) {
configurations[j,colnumber] = meanRanks[meanRanks$config==configurations$config[j],]$normRank
}
}
} else {
cdNemenyi = getNemenyiCD(0.05, length(unique(results$config)), max(results$index))
maxMeanRank = max(meanRanks$meanRank)
minMeanRank = min(meanRanks$meanRank)
# with CD within first results
if( maxMeanRank-minMeanRank<=cdNemenyi ) {
# all on first rank
configurations[, colnumber] = 1
}
else if(maxMeanRank-minMeanRank<=2*cdNemenyi) {
# only two ranks
firstRank = meanRanks$config[meanRanks$meanRank>=maxMeanRank-cdNemenyi]
secondRank = meanRanks$config[meanRanks$meanRank<maxMeanRank-cdNemenyi]
configurations[configurations$config %in% firstRank, colnumber] = 1
configurations[configurations$config %in% secondRank, colnumber] = 1-length(firstRank)/(nrow(meanRanks)-1)
} else {
# three ranks
firstRank <<- meanRanks$config[meanRanks$meanRank>=maxMeanRank-cdNemenyi]
secondRank <<- meanRanks$config[meanRanks$meanRank>minMeanRank+cdNemenyi & meanRanks$meanRank<maxMeanRank-cdNemenyi]
thirdRank <<- meanRanks$config[meanRanks$meanRank<=minMeanRank+cdNemenyi]
configurations[configurations$config %in% firstRank, colnumber] = 1
configurations[configurations$config %in% secondRank, colnumber] = 1-length(firstRank)/(nrow(meanRanks)-1)
configurations[configurations$config %in% thirdRank, colnumber] = 1-(length(firstRank)+length(secondRank))/(nrow(meanRanks)-1)
}
}
colnames(configurations)[colnumber] = paste("FMEAS",dataset,sep="_")
}
else {
skresult = SK(aov( fscore ~ config, results), sig.level=0.05)
colnumber = ncol(configurations)+1
higherRanked = 0
for( j in 1:length(unique(skresult$groups)) ) {
orderedNms = skresult$nms[skresult$ord]
configurations[configurations$config %in% orderedNms[skresult$groups==j], colnumber] = 1-(higherRanked/(nrow(configurations)-1))
higherRanked <- higherRanked+length(orderedNms[skresult$groups==j])
}
colnames(configurations)[colnumber] = paste("FMEAS",dataset,sep="_")
}
colnumber = ncol(configurations)+1
for( j in 1:nrow(configurations) ) {
configurations[j,colnumber] = mean(results[results$config==configurations$config[j],]$fscore, na.rm = TRUE)
}
colnames(configurations)[colnumber] = paste("meanFMEAS",dataset,sep="_")
}
if( "gscore" %in% metricNames ) {
if (NONPARAMETRIC ) {
# Friedman-Nemenyi Test
meanRanks = createMeanRankMat(results, "gscore")
if( PRINTCDCHARTS ) {
printMetricName = "G-measure"
plotObj = plotCDNemenyiFriedman(meanRanks, max(results$index), title=paste("Critical Distance Diagram for ", printMetricName, " and the ", dataset, " data", sep=""))
ggsave(filename = paste(PLOT_PATH,printMetricName,"_",dataset,".png",sep=""), plot=plotObj, device="png", width = CD_EXPORT_WIDTH, height = CD_EXPORT_WIDTH)
}
colnumber = ncol(configurations)+1
if( NONPARAMETRICRANKINGMODE==1 ) {
for( j in 1:nrow(configurations) ) {
if( configurations$config[j] %in% meanRanks$config) {
configurations[j,colnumber] = meanRanks[meanRanks$config==configurations$config[j],]$normRank
}
}
} else {
cdNemenyi = getNemenyiCD(0.05, length(unique(results$config)), max(results$index))
maxMeanRank = max(meanRanks$meanRank)
minMeanRank = min(meanRanks$meanRank)
# with CD within first results
if( maxMeanRank-minMeanRank<=cdNemenyi ) {
# all on first rank
configurations[, colnumber] = 1
}
else if(maxMeanRank-minMeanRank<=2*cdNemenyi) {
# only two ranks
firstRank = meanRanks$config[meanRanks$meanRank>=maxMeanRank-cdNemenyi]
secondRank = meanRanks$config[meanRanks$meanRank<maxMeanRank-cdNemenyi]
configurations[configurations$config %in% firstRank, colnumber] = 1
configurations[configurations$config %in% secondRank, colnumber] = 1-length(firstRank)/(nrow(meanRanks)-1)
} else {
# three ranks
firstRank <<- meanRanks$config[meanRanks$meanRank>=maxMeanRank-cdNemenyi]
secondRank <<- meanRanks$config[meanRanks$meanRank>minMeanRank+cdNemenyi & meanRanks$meanRank<maxMeanRank-cdNemenyi]
thirdRank <<- meanRanks$config[meanRanks$meanRank<=minMeanRank+cdNemenyi]
configurations[configurations$config %in% firstRank, colnumber] = 1
configurations[configurations$config %in% secondRank, colnumber] = 1-length(firstRank)/(nrow(meanRanks)-1)
configurations[configurations$config %in% thirdRank, colnumber] = 1-(length(firstRank)+length(secondRank))/(nrow(meanRanks)-1)
}
}
colnames(configurations)[colnumber] = paste("GMEAS",dataset,sep="_")
}
else {
skresult = SK(aov( gscore ~ config, results), sig.level=0.05)
colnumber = ncol(configurations)+1
higherRanked = 0
for( j in 1:length(unique(skresult$groups)) ) {
orderedNms = skresult$nms[skresult$ord]
configurations[configurations$config %in% orderedNms[skresult$groups==j], colnumber] = 1-(higherRanked/(nrow(configurations)-1))
higherRanked <- higherRanked+length(orderedNms[skresult$groups==j])
}
colnames(configurations)[colnumber] = paste("GMEAS",dataset,sep="_")
}
colnumber = ncol(configurations)+1
for( j in 1:nrow(configurations) ) {
configurations[j,colnumber] = mean(results[results$config==configurations$config[j],]$gscore, na.rm = TRUE)
}
colnames(configurations)[colnumber] = paste("meanGMEAS",dataset,sep="_")
}
if( "mcc" %in% metricNames ) {
if (NONPARAMETRIC ) {
# Friedman-Nemenyi Test
meanRanks = createMeanRankMat(results, "mcc")
if( PRINTCDCHARTS ) {
printMetricName = "MCC"
plotObj = plotCDNemenyiFriedman(meanRanks, max(results$index), title=paste("Critical Distance Diagram for ", printMetricName, " and the ", dataset, " data", sep=""))
ggsave(filename = paste(PLOT_PATH,printMetricName,"_",dataset,".png",sep=""), plot=plotObj, device="png", width = CD_EXPORT_WIDTH, height = CD_EXPORT_WIDTH)
}
colnumber = ncol(configurations)+1
if( NONPARAMETRICRANKINGMODE==1 ) {
for( j in 1:nrow(configurations) ) {
if( configurations$config[j] %in% meanRanks$config) {
configurations[j,colnumber] = meanRanks[meanRanks$config==configurations$config[j],]$normRank
}
}
} else {
cdNemenyi = getNemenyiCD(0.05, length(unique(results$config)), max(results$index))
maxMeanRank = max(meanRanks$meanRank)
minMeanRank = min(meanRanks$meanRank)
# with CD within first results
if( maxMeanRank-minMeanRank<=cdNemenyi ) {
# all on first rank
configurations[, colnumber] = 1
}
else if(maxMeanRank-minMeanRank<=2*cdNemenyi) {
# only two ranks
firstRank = meanRanks$config[meanRanks$meanRank>=maxMeanRank-cdNemenyi]
secondRank = meanRanks$config[meanRanks$meanRank<maxMeanRank-cdNemenyi]
configurations[configurations$config %in% firstRank, colnumber] = 1
configurations[configurations$config %in% secondRank, colnumber] = 1-length(firstRank)/(nrow(meanRanks)-1)
} else {
# three ranks
firstRank <<- meanRanks$config[meanRanks$meanRank>=maxMeanRank-cdNemenyi]
secondRank <<- meanRanks$config[meanRanks$meanRank>minMeanRank+cdNemenyi & meanRanks$meanRank<maxMeanRank-cdNemenyi]
thirdRank <<- meanRanks$config[meanRanks$meanRank<=minMeanRank+cdNemenyi]
configurations[configurations$config %in% firstRank, colnumber] = 1
configurations[configurations$config %in% secondRank, colnumber] = 1-length(firstRank)/(nrow(meanRanks)-1)
configurations[configurations$config %in% thirdRank, colnumber] = 1-(length(firstRank)+length(secondRank))/(nrow(meanRanks)-1)
}
}
colnames(configurations)[colnumber] = paste("MCC",dataset,sep="_")
}
else {
skresult = SK(aov( mcc ~ config, results), sig.level=0.05)
colnumber = ncol(configurations)+1
higherRanked = 0
for( j in 1:length(unique(skresult$groups)) ) {
orderedNms = skresult$nms[skresult$ord]
configurations[configurations$config %in% orderedNms[skresult$groups==j], colnumber] = 1-(higherRanked/(nrow(configurations)-1))
higherRanked <- higherRanked+length(orderedNms[skresult$groups==j])
}
colnames(configurations)[colnumber] = paste("MCC",dataset,sep="_")
}
colnumber = ncol(configurations)+1
for( j in 1:nrow(configurations) ) {
configurations[j,colnumber] = mean(results[results$config==configurations$config[j],]$mcc, na.rm = TRUE)
}
colnames(configurations)[colnumber] = paste("meanMCC",dataset,sep="_")
}
if( CREATEPLOTS ) {
if( "auc" %in% metricNames ) {
plotauc = ggplot(data=results, aes(reorder(config, auc, FUN=mean), auc)) +
ggtitle(paste("AUC ordered by the mean value for", dataset, "data")) +
geom_boxplot() + stat_summary(fun.y=mean, geom="point", size=3, shape=9) +
scale_y_continuous(limits=c(0,1.0)) + coord_flip() +
theme(axis.text=element_text(size=8), axis.title=element_text(size=8), title=element_text(size=8), axis.text.x = element_text(angle = 90, hjust = 1))
ggsave(filename=paste(PLOT_PATH,"AUC_",dataset,".pdf",sep=""), plot=plotauc)
}
if( "fscore" %in% metricNames ) {
plotfscore = ggplot(data=results, aes(reorder(config, fscore, FUN=mean), fscore)) +
ggtitle(paste("F-measure ordered by the mean value for", dataset, "data")) +
geom_boxplot() + stat_summary(fun.y=mean, geom="point", size=3, shape=9) +
scale_y_continuous(limits=c(0,1.0)) + coord_flip() +
theme(axis.text=element_text(size=8), axis.title=element_text(size=8), title=element_text(size=8), axis.text.x = element_text(angle = 90, hjust = 1))
ggsave(filename=paste(PLOT_PATH,"FMEASURE_",dataset,".pdf",sep=""), plot=plotfscore)
}
if( "gscore" %in% metricNames ) {
plotgscore = ggplot(data=results, aes(reorder(config, gscore, FUN=mean), gscore)) +
ggtitle(paste("G-measure ordered by the mean value for", dataset, "data")) +
geom_boxplot() + stat_summary(fun.y=mean, geom="point", size=3, shape=9) +
scale_y_continuous(limits=c(0,1.0)) + coord_flip() +
theme(axis.text=element_text(size=8), axis.title=element_text(size=8), title=element_text(size=8), axis.text.x = element_text(angle = 90, hjust = 1))
ggsave(filename=paste(PLOT_PATH,"GMEASURE_",dataset,".pdf",sep=""), plot=plotgscore)
}
if( "mcc" %in% metricNames ) {
plotmcc = ggplot(data=results, aes(reorder(config, mcc, FUN=mean), mcc)) +
ggtitle(paste("MCC ordered by the mean value for", dataset, "data")) +
geom_boxplot() + stat_summary(fun.y=mean, geom="point", size=3, shape=9) +
scale_y_continuous(limits=c(0,1.0)) + coord_flip() +
theme(axis.text=element_text(size=8), axis.title=element_text(size=8), title=element_text(size=8), axis.text.x = element_text(angle = 90, hjust = 1))
ggsave(filename=paste(PLOT_PATH,"MCC_",dataset,".pdf",sep=""), plot=plotmcc)
}
}
}
dbDisconnect(mydb)
skresults = configurations
rownames(skresults) = skresults$config
skresults$config = NULL
skresults$MEANRANK = rowMeans(skresults[!startsWith(colnames(skresults), "mean")], na.rm=TRUE)
skresults$config = as.factor(rownames(skresults))
for( i in 1:nrow(skresults) ) {
skresults$approach[i] = strsplit(rownames(skresults)[[i]],"-")[[1]][1]
splitLength = nchar(strsplit(rownames(skresults)[[i]],"-")[[1]][1])
skresults$classifier[i] = substring(rownames(skresults)[[i]],splitLength+2)
}
return(skresults)
}
plotBestResults = function(results, rq, metricsString) {
best = results
approaches = unique(best$approach)
for( j in 1:length(approaches)) {
maxval = max(best[best$approach==approaches[j],]$MEANRANK)
for( i in nrow(best):1 ) {
if( best$approach[i]==approaches[j] && best$MEANRANK[i]<maxval ) {
best = best[-i,]
}
}
}
best$MEANRANK = round(best$MEANRANK, digits=3)
textsize = 12
plotBest = ggplot(data=best, aes(reorder(config, MEANRANK, FUN=mean), MEANRANK)) +
ggtitle(paste("Ranking of approaches using",metricsString)) + ylab("Mean rankscore") + xlab("Approach") +
geom_bar(stat="identity" ) + geom_label(aes(label=MEANRANK), hjust=0, nudge_y = 0.002) +
scale_y_continuous(limits=c(0,1.02), breaks=c(0,0.25,0.5,0.75,1)) + coord_flip() +
theme(axis.text=element_text(size=textsize), axis.title=element_text(size=textsize), title=element_text(size=textsize))
print(plotBest)
ggsave(filename=paste(PLOT_PATH,"BEST_", rq, ".pdf", sep=""), plot=plotBest)
return(best)
}
writeResultsTableRQ1 = function(rq1best, rq4results, rq5results) {
resultsTableFrame = round(rq1best[,1:41], digits=2)
colnumber = ncol(resultsTableFrame)+1
resultsTableFrame[,colnumber] = paste(resultsTableFrame$meanAUC_JURECZKO, " (", resultsTableFrame$AUC_JURECZKO, ")", sep="")
colnames(resultsTableFrame)[colnumber] = "auc"
colnumber = ncol(resultsTableFrame)+1
resultsTableFrame[,colnumber] = paste(resultsTableFrame$meanFMEAS_JURECZKO, " (", resultsTableFrame$FMEAS_JURECZKO, ")", sep="")
colnames(resultsTableFrame)[colnumber] = "F-measure"
colnumber = ncol(resultsTableFrame)+1
resultsTableFrame[,colnumber] = paste(resultsTableFrame$meanGMEAS_JURECZKO, " (", resultsTableFrame$GMEAS_JURECZKO, ")", sep="")
colnames(resultsTableFrame)[colnumber] = "G-measure"
colnumber = ncol(resultsTableFrame)+1
resultsTableFrame[,colnumber] = paste(resultsTableFrame$meanMCC_JURECZKO, " (", resultsTableFrame$MCC_JURECZKO, ")", sep="")
colnames(resultsTableFrame)[colnumber] = "MCC"
colnumber = ncol(resultsTableFrame)+1
colnumber = ncol(resultsTableFrame)+1
resultsTableFrame[,colnumber] = paste(round(rq4results[rq4results$config %in% rownames(resultsTableFrame),]$meanAUC_FILTERJURECZKO-rq1best$meanAUC_JURECZKO, digits=2), " / " ,round(rq5results[rq5results$config %in% rownames(resultsTableFrame),]$meanAUC_SELECTEDJURECZKO-rq1best$meanAUC_JURECZKO, digits=2), sep="")
colnames(resultsTableFrame)[colnumber] = "auc"
colnumber = ncol(resultsTableFrame)+1
resultsTableFrame[,colnumber] = paste(round(rq4results[rq4results$config %in% rownames(resultsTableFrame),]$meanFMEAS_FILTERJURECZKO-rq1best$meanFMEAS_JURECZKO, digits=2), " / " ,round(rq5results[rq5results$config %in% rownames(resultsTableFrame),]$meanFMEAS_SELECTEDJURECZKO-rq1best$meanFMEAS_JURECZKO, digits=2), sep="")
colnames(resultsTableFrame)[colnumber] = "F-measure"
colnumber = ncol(resultsTableFrame)+1
resultsTableFrame[,colnumber] = paste(round(rq4results[rq4results$config %in% rownames(resultsTableFrame),]$meanGMEAS_FILTERJURECZKO-rq1best$meanGMEAS_JURECZKO, digits=2), " / " ,round(rq5results[rq5results$config %in% rownames(resultsTableFrame),]$meanGMEAS_SELECTEDJURECZKO-rq1best$meanGMEAS_JURECZKO, digits=2), sep="")
colnames(resultsTableFrame)[colnumber] = "G-measure"
colnumber = ncol(resultsTableFrame)+1
resultsTableFrame[,colnumber] = paste(round(rq4results[rq4results$config %in% rownames(resultsTableFrame),]$meanMCC_FILTERJURECZKO-rq1best$meanMCC_JURECZKO, digits=2), " / " ,round(rq5results[rq5results$config %in% rownames(resultsTableFrame),]$meanMCC_SELECTEDJURECZKO-rq1best$meanMCC_JURECZKO, digits=2), sep="")
colnames(resultsTableFrame)[colnumber] = "MCC"
colnumber = ncol(resultsTableFrame)+1
colnumber = ncol(resultsTableFrame)+1
resultsTableFrame[,colnumber] = paste(resultsTableFrame$meanAUC_MDP, " (", resultsTableFrame$AUC_MDP, ")", sep="")
colnames(resultsTableFrame)[colnumber] = "auc"
colnumber = ncol(resultsTableFrame)+1
resultsTableFrame[,colnumber] = paste(resultsTableFrame$meanFMEAS_MDP, " (", resultsTableFrame$FMEAS_MDP, ")", sep="")
colnames(resultsTableFrame)[colnumber] = "F-measure"
colnumber = ncol(resultsTableFrame)+1
resultsTableFrame[,colnumber] = paste(resultsTableFrame$meanGMEAS_MDP, " (", resultsTableFrame$GMEAS_MDP, ")", sep="")
colnames(resultsTableFrame)[colnumber] = "G-measure"
colnumber = ncol(resultsTableFrame)+1
resultsTableFrame[,colnumber] = paste(resultsTableFrame$meanMCC_MDP, " (", resultsTableFrame$MCC_MDP, ")", sep="")
colnames(resultsTableFrame)[colnumber] = "MCC"
colnumber = ncol(resultsTableFrame)+1
colnumber = ncol(resultsTableFrame)+1
resultsTableFrame[,colnumber] = paste(resultsTableFrame$meanAUC_AEEEM_LDHHWCHU, " (", resultsTableFrame$AUC_AEEEM_LDHHWCHU, ")", sep="")
colnames(resultsTableFrame)[colnumber] = "auc"
colnumber = ncol(resultsTableFrame)+1
resultsTableFrame[,colnumber] = paste(resultsTableFrame$meanFMEAS_AEEEM_LDHHWCHU, " (", resultsTableFrame$FMEAS_AEEEM_LDHHWCHU, ")", sep="")
colnames(resultsTableFrame)[colnumber] = "F-measure"
colnumber = ncol(resultsTableFrame)+1
resultsTableFrame[,colnumber] = paste(resultsTableFrame$meanGMEAS_AEEEM_LDHHWCHU, " (", resultsTableFrame$GMEAS_AEEEM_LDHHWCHU, ")", sep="")
colnames(resultsTableFrame)[colnumber] = "G-measure"
colnumber = ncol(resultsTableFrame)+1
resultsTableFrame[,colnumber] = paste(resultsTableFrame$meanMCC_AEEEM_LDHHWCHU, " (", resultsTableFrame$MCC_AEEEM_LDHHWCHU, ")", sep="")
colnames(resultsTableFrame)[colnumber] = "MCC"
colnumber = ncol(resultsTableFrame)+1
colnumber = ncol(resultsTableFrame)+1
resultsTableFrame[,colnumber] = paste(resultsTableFrame$meanAUC_NETGENE, " (", resultsTableFrame$AUC_NETGENE, ")", sep="")
colnames(resultsTableFrame)[colnumber] = "auc"
colnumber = ncol(resultsTableFrame)+1
resultsTableFrame[,colnumber] = paste(resultsTableFrame$meanFMEAS_NETGENE, " (", resultsTableFrame$FMEAS_NETGENE, ")", sep="")
colnames(resultsTableFrame)[colnumber] = "F-measure"
colnumber = ncol(resultsTableFrame)+1
resultsTableFrame[,colnumber] = paste(resultsTableFrame$meanGMEAS_NETGENE, " (", resultsTableFrame$GMEAS_NETGENE, ")", sep="")
colnames(resultsTableFrame)[colnumber] = "G-measure"
colnumber = ncol(resultsTableFrame)+1
resultsTableFrame[,colnumber] = paste(resultsTableFrame$meanMCC_NETGENE, " (", resultsTableFrame$MCC_NETGENE, ")", sep="")
colnames(resultsTableFrame)[colnumber] = "MCC"
colnumber = ncol(resultsTableFrame)+1
colnumber = ncol(resultsTableFrame)+1
resultsTableFrame[,colnumber] = paste(resultsTableFrame$meanAUC_RELINK, " (", resultsTableFrame$AUC_RELINK, ")", sep="")
colnames(resultsTableFrame)[colnumber] = "auc"
colnumber = ncol(resultsTableFrame)+1
resultsTableFrame[,colnumber] = paste(resultsTableFrame$meanFMEAS_RELINK, " (", resultsTableFrame$FMEAS_RELINK, ")", sep="")
colnames(resultsTableFrame)[colnumber] = "F-measure"
colnumber = ncol(resultsTableFrame)+1
resultsTableFrame[,colnumber] = paste(resultsTableFrame$meanGMEAS_RELINK, " (", resultsTableFrame$GMEAS_RELINK, ")", sep="")
colnames(resultsTableFrame)[colnumber] = "G-measure"
colnumber = ncol(resultsTableFrame)+1
resultsTableFrame[,colnumber] = paste(resultsTableFrame$meanMCC_RELINK, " (", resultsTableFrame$MCC_RELINK, ")", sep="")
colnames(resultsTableFrame)[colnumber] = "MCC"
colnumber = ncol(resultsTableFrame)+1
tableCaption = "Mean results over all products with rankscores in brackets. Bold-faced values are top-ranking for the metric on the data set. For FILTERJURECZKO and SELECTEDJURECKO, we show the difference in the mean values to JURECZKO."
tablePart1 = print.xtable(xtable(resultsTableFrame[,42:53]))
tablePart2 = print.xtable(xtable(resultsTableFrame[,54:65]))
tablePart1 = sub("\\centering", "", tablePart1, fixed=TRUE)
tablePart1 = sub("\\begin{table}[ht]", "\\begin{table*}\n\\scriptsize\\centering\\begin{sideways}", tablePart1, fixed=TRUE)
tablePart1 = sub("rllllrrrrllll", "|r|llll|llll|llll|", tablePart1, fixed=TRUE)
tablePart1 = sub("\\hline", "\\hline\n & \\multicolumn{4}{c|}{JURECZKO} & \\multicolumn{4}{c|}{FILTERJURECZKO / SELECTEDJURECZKO} & \\multicolumn{4}{c|}{MDP} \\\\\n\\hline", tablePart1, fixed=TRUE)
tablePart1 = sub("\\end{tabular}\n\\end{table}\n", "", tablePart1, fixed=TRUE)
tablePart2 = sub("\\begin{table}[ht]\n\\centering\n\\begin{tabular}{rllllllllllll}\n \\hline\n", "\\hline\n & \\multicolumn{4}{c|}{AEEEM} & \\multicolumn{4}{c|}{NETGENE} & \\multicolumn{4}{c|}{RELINK} \\\\\n\\hline", tablePart2, fixed=TRUE)
tablePart2 = sub(".*\\\\hline\\\n & \\\\\\multicolumn","\\\\hline\n & \\\\multicolumn",tablePart2, fixed=FALSE)
tablePart2 = sub("\\end{table}", paste("\\end{sideways}\n\\caption{", tableCaption, "}\n\\label{tbl:results}\n\\end{table*}", sep=""), tablePart2, fixed=TRUE)
tableStr = paste(tablePart1, tablePart2, sep="")
tableStr = gsub("NaN (NA)", "-", tableStr, fixed=TRUE)
tableStr = gsub("NaN / NaN", "-", tableStr, fixed=TRUE)
tableStr = gsub("& &", "& - &", tableStr, fixed=TRUE)
tableStr = gsub("0 ", "0.00 ", tableStr, fixed=TRUE)
tableStr = gsub("auc.1", "\\emph{AUC}", tableStr, fixed=TRUE)
tableStr = gsub("auc.2", "\\emph{AUC}", tableStr, fixed=TRUE)
tableStr = gsub("auc.3", "\\emph{AUC}", tableStr, fixed=TRUE)
tableStr = gsub("auc.4", "\\emph{AUC}", tableStr, fixed=TRUE)
tableStr = gsub("auc.5", "\\emph{AUC}", tableStr, fixed=TRUE)
tableStr = gsub("auc ", "\\emph{AUC} ", tableStr, fixed=TRUE)
tableStr = gsub("F-measure.1", "\\emph{F-measure}", tableStr, fixed=TRUE)
tableStr = gsub("F-measure.2", "\\emph{F-measure}", tableStr, fixed=TRUE)
tableStr = gsub("F-measure.3", "\\emph{F-measure}", tableStr, fixed=TRUE)
tableStr = gsub("F-measure.4", "\\emph{F-measure}", tableStr, fixed=TRUE)
tableStr = gsub("F-measure.5", "\\emph{F-measure}", tableStr, fixed=TRUE)
tableStr = gsub("F-measure ", "\\emph{F-measure} ", tableStr, fixed=TRUE)
tableStr = gsub("G-measure.1", "\\emph{G-measure}", tableStr, fixed=TRUE)
tableStr = gsub("G-measure.2", "\\emph{G-measure}", tableStr, fixed=TRUE)
tableStr = gsub("G-measure.3", "\\emph{G-measure}", tableStr, fixed=TRUE)
tableStr = gsub("G-measure.4", "\\emph{G-measure}", tableStr, fixed=TRUE)
tableStr = gsub("G-measure.5", "\\emph{G-measure}", tableStr, fixed=TRUE)
tableStr = gsub("G-measure ", "\\emph{G-measure} ", tableStr, fixed=TRUE)
tableStr = gsub("MCC.1", "\\emph{MCC}", tableStr, fixed=TRUE)
tableStr = gsub("MCC.2", "\\emph{MCC}", tableStr, fixed=TRUE)
tableStr = gsub("MCC.3", "\\emph{MCC}", tableStr, fixed=TRUE)
tableStr = gsub("MCC.4", "\\emph{MCC}", tableStr, fixed=TRUE)
tableStr = gsub("MCC.5", "\\emph{MCC}", tableStr, fixed=TRUE)
tableStr = gsub("MCC ", "\\emph{MCC} ", tableStr, fixed=TRUE)
tableStr = gsub("([0-9]\\.[0-9][0-9]? \\(1\\))", "\\\\textbf\\{\\1\\}", tableStr, perl=TRUE )
write(tableStr, file=paste(TABLES_PATH, "resultsTable.tex", sep=""))
}
compareDatasets = function(set1, set2, metric) {
mydb = dbConnect(MySQL(), user=MYSQL_USER, password=MYSQL_PASSWORT, host=MYSQL_HOST, port=MYSQL_PORT, dbname=MYSQL_DBNAME)
sqlStatement = paste("SELECT q1.config, q1.mean1, q2.mean2 FROM ",
"(SELECT concat(substring(configurationName, ",nchar(set1)+2,"), '-', classifier) as config, avg(", metric,
") as mean1 FROM resultsView WHERE configurationName LIKE '",set1, "-%' GROUP BY configurationName, classifier) q1, ",
"(SELECT concat(substring(configurationName, ",nchar(set2)+2,"), '-', classifier) as config, avg(", metric,
") as mean2 FROM resultsView WHERE configurationName LIKE '",set2, "-%' GROUP BY configurationName, classifier) q2 ",
"WHERE q1.config=q2.config", sep = "")
rs = dbSendQuery(mydb, sqlStatement)
results = fetch(rs, n=-1)
cat(paste("metric:", metric, "\n"))
cat(paste("mean", set1, ":", mean(results$mean1), "\n"))
cat(paste("mean", set2, ":", mean(results$mean2)), "\n")
cat(paste("mean(abs(mean1-mean2)):", mean(abs(results$mean1-results$mean2)), "\n"))
cat(paste("sd(abs(mean1-mean2)):", sd(abs(results$mean1-results$mean2)), "\n"))
print(wilcox.test(results$mean1, results$mean2))
dbDisconnect(mydb)
}
evalRQ2 = function() {
mydb = dbConnect(MySQL(), user=MYSQL_USER, password=MYSQL_PASSWORT, host=MYSQL_HOST, port=MYSQL_PORT, dbname=MYSQL_DBNAME)
sqlStatement = "SELECT count1, count2, count1/count2 FROM (SELECT count(*) as count1 FROM results WHERE configurationName NOT LIKE 'F%' AND configurationName NOT LIKE 'S%' AND recall>=0.75 AND results.precision>=0.75 AND error<=0.25) q1, (SELECT count(*) as count2 FROM results WHERE configurationName NOT LIKE 'F%' AND configurationName NOT LIKE 'S%') q2;"
rs = dbSendQuery(mydb, sqlStatement)
results = fetch(rs, n=-1)
cat(paste("results with baselines for Zimmermann\n"))
cat(paste("number of single results that fulfill the criterion: ", results[1,1], "\n"))
cat(paste("number of single results that do not fulfill the criterion:", results[1,2], "\n"))
cat(paste("rate of fulfilling the criterion: ", results[1,3], "\n"))
cat("\n")
sqlStatement = "SELECT count1, count2, count1/count2 FROM (SELECT count(*) as count1 FROM results WHERE configurationName NOT LIKE 'F%' AND configurationName NOT LIKE 'S%' and configurationName NOT LIKE '%CV' AND classifier!='FIX' AND classifier!='RANDOM' AND recall>=0.75 AND results.precision>=0.75 AND error<=0.25) q1, (SELECT count(*) as count2 FROM results WHERE configurationName NOT LIKE 'F%' AND configurationName NOT LIKE 'S%' and configurationName NOT LIKE '%CV' AND classifier!='FIX' AND classifier!='RANDOM') q2;"
rs = dbSendQuery(mydb, sqlStatement)
results = fetch(rs, n=-1)
cat(paste("results without baselines for Zimmermann\n"))
cat(paste("number of single results that fulfill the criterion: ", results[1,1], "\n"))
cat(paste("number of single results that do not fulfill the criterion:", results[1,2], "\n"))
cat(paste("rate of fulfilling the criterion: ", results[1,3], "\n"))
cat("\n")
sqlStatement = "SELECT count1, count2, count1/count2 FROM (SELECT count(*) as count1 FROM results WHERE configurationName NOT LIKE 'F%' AND configurationName NOT LIKE 'S%' AND recall>=0.70 AND results.precision>=0.5) q1, (SELECT count(*) as count2 FROM results WHERE configurationName NOT LIKE 'F%' AND configurationName NOT LIKE 'S%') q2;"
rs = dbSendQuery(mydb, sqlStatement)
results = fetch(rs, n=-1)
cat(paste("results with baselines for He\n"))
cat(paste("number of single results that fulfill the criterion: ", results[1,1], "\n"))
cat(paste("number of single results that do not fulfill the criterion:", results[1,2], "\n"))
cat(paste("rate of fulfilling the criterion: ", results[1,3], "\n"))
cat("\n")
sqlStatement = "SELECT count1, count2, count1/count2 FROM (SELECT count(*) as count1 FROM results WHERE configurationName NOT LIKE 'F%' AND configurationName NOT LIKE 'S%' and configurationName NOT LIKE '%CV' AND classifier!='FIX' AND classifier!='RANDOM' AND recall>=0.7 AND results.precision>=0.5) q1, (SELECT count(*) as count2 FROM results WHERE configurationName NOT LIKE 'F%' AND configurationName NOT LIKE 'S%' and configurationName NOT LIKE '%CV' AND classifier!='FIX' AND classifier!='RANDOM') q2;"
rs = dbSendQuery(mydb, sqlStatement)
results = fetch(rs, n=-1)
cat(paste("results without baselines for He\n"))
cat(paste("number of single results that fulfill the criterion: ", results[1,1], "\n"))
cat(paste("number of single results that do not fulfill the criterion:", results[1,2], "\n"))
cat(paste("rate of fulfilling the criterion: ", results[1,3], "\n"))
cat("\n")
sqlStatement = "SELECT res1.configurationName, res1.classifier, cnt1 as cnt, cnt1/cnt2 as rate FROM (SELECT configurationName, classifier, count(*) as cnt1 FROM results WHERE configurationName NOT LIKE 'F%' AND recall>=0.75 AND results.precision>=0.75 AND error<=0.25 GROUP BY configurationName, classifier) as res1, (SELECT configurationName, classifier, count(*) as cnt2 FROM results WHERE configurationName NOT LIKE 'F%' GROUP BY configurationName, classifier) as res2 WHERE res1.configurationName=res2.configurationName AND res1.classifier=res2.classifier ORDER BY rate DESC;"
rs = dbSendQuery(mydb, sqlStatement)
results = fetch(rs, n=-1)
cat("distinct products with at least 100 classes that fulfill the criterion:\n")
cat("configuration\t\t\tclassifier\t\tcount\trate\n")
for( i in 1:nrow(results) ) {
cat(paste(str_pad(results[i,1], width=25, side="right"), str_pad(results[i,2], width=20, side="right"), results[i,3], results[i,4], "\n", sep="\t"))
}
cat("\n")
sqlStatement = "SELECT distinct productName FROM results WHERE configurationName NOT LIKE 'F%' AND configurationName NOT LIKE 'S%' AND recall>=0.75 AND results.precision>=0.75 AND error<=0.25 AND testsize>=100 GROUP BY configurationName, classifier"
rs = dbSendQuery(mydb, sqlStatement)
results = fetch(rs, n=-1)
cat("distinct products with at least 100 classes that fulfill the criterion:\n")
cat(paste(results[,1], collapse = "\n"))
cat("\n\n")
sqlStatement = "SELECT distinct productName FROM results WHERE configurationName NOT LIKE 'F%' AND configurationName NOT LIKE 'S%' AND recall>=0.75 AND results.precision>=0.75 AND error<=0.25 AND testsize<100 GROUP BY configurationName, classifier"
rs = dbSendQuery(mydb, sqlStatement)
results = fetch(rs, n=-1)
cat("distinct products with less than 100 classes that fulfill the criterion:\n")
cat(paste(results[,1], collapse = "\n"))
cat("\n\n")
sqlStatement = "SELECT productName as Product, testsize, sum(cnt1) as cnt, prec FROM (SELECT productName, testsize, if(count(*)=1, 1, count(*)/10) as cnt1, 1.0 as recall, (tp+fn)/testsize as 'prec' FROM results WHERE configurationName NOT LIKE 'F%' AND configurationName NOT LIKE 'S%' AND configurationName NOT LIKE '%CV' AND classifier!='FIX' AND classifier!='RANDOM' AND recall>=0.75 AND results.precision>=0.75 AND error<=0.25 GROUP BY configurationName, classifier, productName) as res GROUP BY productName, testsize UNION SELECT DISTINCT productName as Product, testsize, 0 as cnt, (tp+fn)/testsize as prec FROM results WHERE classifier='FIX' and (tp+fn)/testsize>=0.75 and productName!='pbeans1.csv' ORDER BY Product"
rs = dbSendQuery(mydb, sqlStatement)
results = fetch(rs, n=-1)
cat("distinct products with less than 100 classes that fulfill the criterion:\n")
cat(paste(results[,1], collapse = "\n"))
cat("\n\n")
rownames(results) = results$Product
results$Product = NULL
results$Data = 'JURECZKO'
results = results[,c(4,1,2,3)]
captionStr = "Products where any \\ac{CPDP} approach fulfills the criterion by Zimmermann\\etal~\\cite{Zimmermann2009} and the \\emph{precision} of the trivial prediction FIX."
tableStr = print.xtable(xtable(results, caption=captionStr))
#tableStr = sub("\\begin{table}", "\\begin{table*}", tableStr, fixed=TRUE)
tableStr = sub("rlrrr", "|r|l|l|l|l|", tableStr, fixed=TRUE)
tableStr = sub("\\hline\n", "\\hline\n \\textbf{Product}", tableStr, fixed=TRUE)
tableStr = sub("Data", "\\textbf{Data Set}", tableStr, fixed=TRUE)
tableStr = sub("testsize", "\\textbf{\\#Inst.}", tableStr, fixed=TRUE)
tableStr = sub("cnt", "\\textbf{\\#Appr.}", tableStr, fixed=TRUE)
tableStr = sub("prec \\\\ \n", "\\emph{precision} \\\\ &&&& FIX \\\\ \\hline\n", tableStr, fixed=TRUE)
tableStr = gsub(".csv", "", tableStr, fixed=TRUE)
tableStr = sub("openintents & JURECZKO", "openintents & RELINK", tableStr, fixed=TRUE)
tableStr = sub("\\end{table}", "\\label{tbl:productsSuccess}\n\\end{table}", tableStr, fixed=TRUE)
write(tableStr, paste(TABLES_PATH,"productsSuccess.tex", sep=""))
dbDisconnect(mydb)
}
getNemenyiCD <- function(pval, nConfigs, nData) {
return(qtukey(1 - pval, nConfigs, Inf) / sqrt((nConfigs*(nConfigs+1))/(12*nData)))
}
plotCDNemenyiFriedman <- function(meanRanks, nData, pval=0.05, title="" ) {
nConfigs = nrow(meanRanks)
cdNemenyi = getNemenyiCD(pval, nConfigs, nData)
roundedMaxMeanRank = ceiling(max(meanRanks$meanRank))
roundedMinMeanRank = floor(min(meanRanks$meanRank))
maxMeanRank = max(meanRanks$meanRank)
minMeanRank = min(meanRanks$meanRank)
halfEntries = ceiling(nrow(meanRanks)/2)
red = 0.8
green = 0
stepsize = 1.6/max(meanRanks$rank)
colors = list()
for( j in 1:max(meanRanks$rank) ) {
colors[j] = rgb(green, red, 0)
if( green+stepsize>0.8 ) {
green = 0.8
red = red-stepsize
if( red<0 ) {
red = 0
}
} else {
green = green+stepsize
if( green>0.8 ) {
green = 0.8
}
}
}
meanRanks <- meanRanks[order(meanRanks$meanRank, decreasing = FALSE),]
g <- ggplot(data=meanRanks)
# create lines for configurations
for( j in 1:nrow(meanRanks)) {
xval = meanRanks$meanRank[j]
if( j<=halfEntries ) {
yend = j
xlinestart = roundedMinMeanRank
xtext = roundedMinMeanRank-0.2/roundedMaxMeanRank
hjust = "right"
} else {
yend = j-2*(j-halfEntries)
xlinestart= roundedMaxMeanRank
xtext = roundedMaxMeanRank+0.2/roundedMaxMeanRank
hjust="left"
}
if( meanRanks$meanRank[j] <= minMeanRank+cdNemenyi ) {
color = "#00cc00"
}
else if( meanRanks$meanRank[j] >= maxMeanRank-cdNemenyi ) {
color = "#cc0000"
}
else {
color = "#0000cc"
}
color = colors[[meanRanks$rank[j]]]
g <- g + geom_segment(x=xval, y=0, xend=xval, yend=-yend, color = color)
g <- g + geom_segment(x=xlinestart, y=-yend, xend=xval, yend=-yend, color = color)
g <- g + geom_text(label=meanRanks$config[j], x=xtext, y=-yend, hjust=hjust, color=color)
}
# create axis
if( roundedMaxMeanRank>100 ) {
breaks = 1:(roundedMaxMeanRank/5)*5
} else if( roundedMaxMeanRank>80 ) {
breaks = 1:(roundedMaxMeanRank/4)*4
} else if( roundedMaxMeanRank>60 ) {
breaks = 1:(roundedMaxMeanRank/3)*3
} else if( roundedMaxMeanRank>40 ) {
breaks = 1:(roundedMaxMeanRank/2)*2
} else {
breaks = 1:roundedMaxMeanRank
}
breaks <<- breaks
g <- g + geom_segment(x=0,y=0,xend=roundedMaxMeanRank,yend=0)
g <- g + geom_text(label="Mean Rank", x=roundedMinMeanRank-0.2/roundedMaxMeanRank, y=0.01*halfEntries, hjust="right", vjust="bottom")
for( j in roundedMinMeanRank:roundedMaxMeanRank ) {
if( j %in% breaks ) {
g <- g + geom_segment(x=j,y=0,xend=j,yend=0.01*halfEntries)
g <- g + geom_text(label=j, x=j, y=0.015*halfEntries, vjust="bottom")
}
}
# add critical distance
critDistY = 3
g <- g + geom_segment(x=minMeanRank,y=critDistY,xend=minMeanRank+cdNemenyi, yend=critDistY)
g <- g + geom_segment(x=minMeanRank+cdNemenyi,y=critDistY-0.005*halfEntries, xend=minMeanRank+cdNemenyi, yend=critDistY+0.005*halfEntries)
g <- g + geom_segment(x=minMeanRank,y=critDistY-0.005*halfEntries, xend=minMeanRank, yend=critDistY+0.005*halfEntries)
g <- g + geom_text(label=paste("Critical Distance =", round(cdNemenyi, digits=3)), x=minMeanRank-0.2/roundedMaxMeanRank, y=critDistY, hjust="right")
# set scales
g <- g + scale_y_continuous(limits=c(-halfEntries,1), breaks=NULL, labels=NULL)
g <- g + scale_x_continuous(limits=c(roundedMinMeanRank-1,roundedMaxMeanRank+1), breaks=breaks, labels=NULL)
g <- g + theme(axis.ticks.x = element_blank())
# set title
g <- g + ggtitle(title)
g <- g + coord_cartesian(xlim=c(roundedMinMeanRank-0.2*roundedMaxMeanRank,roundedMaxMeanRank+0.2*roundedMaxMeanRank))
return(g)
}
createMeanRankMat <- function(results, column) {
nemenyi.groups = factor(as.character(results$config))
nemenyi.blocks = factor(results$index)
nemenyi.y = results[[column]]
nemenyi.n = length(levels(nemenyi.blocks))
nemenyi.k = length(levels(nemenyi.groups))
nemenyi.y = nemenyi.y[order(nemenyi.groups, nemenyi.blocks)]
nemenyi.mat = matrix(nemenyi.y, nrow = nemenyi.n, ncol = nemenyi.k, byrow = FALSE)
for (nemenyi.i in 1:length(nemenyi.mat[, 1])) nemenyi.mat[nemenyi.i, ] <- rank(nemenyi.mat[nemenyi.i, ])
nemenyi.mnsum = data.frame(meanRank=colMeans(nemenyi.mat))
nemenyi.mnsum$config = levels(nemenyi.groups)
# create ranks
# break if difference > cd
nemenyi.cd = getNemenyiCD(0.05, length(unique(results$config)), max(results$index))
nemenyi.mnsum = nemenyi.mnsum[order(-nemenyi.mnsum$meanRank),]
currentRank = 1
nemenyi.mnsum$rank[1] = 1
for( i in 2:nrow(nemenyi.mnsum )) {
if(nemenyi.mnsum$meanRank[i-1]-nemenyi.mnsum$meanRank[i]>nemenyi.cd) {
currentRank = currentRank+1
}
nemenyi.mnsum$rank[i] = currentRank
}
nemenyi.mnsum$normRank = 1-(nemenyi.mnsum$rank-1)/(max(nemenyi.mnsum$rank)-1)
return(nemenyi.mnsum)
}
#####################
# Result evaluation #
#####################
# Performance metrics used
# Script can only handle auc, fscore, gscore, and mcc
# To use other metrics available in the MySQL database,
# the functions above must be modified, because we do
# some string matching with the metric names.
metricNamesRQ1 = c("auc","fscore","gscore","mcc")
# Datasets that are used for different research questions.
datasetsRQ1 = c("JURECZKO", "MDP", "AEEEM_LDHHWCHU", "RELINK", "NETGENE")
datasetsRQ3 = c("FILTERJURECZKO")
datasetsRQ4 = c("SELECTEDJURECZKO")
rq1results = evaluateCPDPBenchmark(metricNamesRQ1, datasetsRQ1)
rq1best = plotBestResults(rq1results, "RQ1", "AUC, F-measure, G-Measure, and MCC")
rq3results = evaluateCPDPBenchmark(metricNamesRQ1, datasetsRQ3)
rq4results = evaluateCPDPBenchmark(metricNamesRQ1, datasetsRQ4)
writeResultsTableRQ1(rq1best, rq3results, rq4results)
evalRQ2()
cat("comparing JURECZKO and FILTERJURECZKO\n")
compareDatasets("JURECZKO", "FILTERJURECZKO", "auc")
compareDatasets("JURECZKO", "FILTERJURECZKO", "fscore")
compareDatasets("JURECZKO", "FILTERJURECZKO", "gscore")
compareDatasets("JURECZKO", "FILTERJURECZKO", "mcc")
cat("comparing JURECZKO and SELECTEDJURECZKO\n")
compareDatasets("JURECZKO", "SELECTEDJURECZKO", "auc")
compareDatasets("JURECZKO", "SELECTEDJURECZKO", "fscore")
compareDatasets("JURECZKO", "SELECTEDJURECZKO", "gscore")
compareDatasets("JURECZKO", "SELECTEDJURECZKO", "mcc")
cat("comparing AEEEM and AEEEM_LDHH")
compareDatasets("AEEEM", "AEEEM_LDHH", "auc")
compareDatasets("AEEEM", "AEEEM_LDHH", "fscore")
compareDatasets("AEEEM", "AEEEM_LDHH", "gscore")
compareDatasets("AEEEM", "AEEEM_LDHH", "mcc")
cat("comparing AEEEM and AEEEM_WCHU")
compareDatasets("AEEEM", "AEEEM_WCHU", "auc")
compareDatasets("AEEEM", "AEEEM_WCHU", "fscore")
compareDatasets("AEEEM", "AEEEM_WCHU", "gscore")
compareDatasets("AEEEM", "AEEEM_WCHU", "mcc")
cat("comparing AEEEM and AEEEM_LDHHWCHU")
compareDatasets("AEEEM", "AEEEM_LDHHWCHU", "auc")
compareDatasets("AEEEM", "AEEEM_LDHHWCHU", "fscore")
compareDatasets("AEEEM", "AEEEM_LDHHWCHU", "gscore")
compareDatasets("AEEEM", "AEEEM_LDHHWCHU", "mcc")