Difference between revisions of "2008:Audio Tag Classification Results"

From MIREX Wiki
(CSV Files Without Rounding)
m (Robot: Automated text replacement (-/mirex/2008/results/ +/mirex/results/2008/))
 
(27 intermediate revisions by 6 users not shown)
Line 1: Line 1:
 
==Introduction==
 
==Introduction==
These are the results for the 2008 running of the Audio Tag Classification task. For background information about this task set please refer to the [[Audio Tag Classification]] page.  
+
These are the results for the 2008 running of the Audio Tag Classification task. For background information about this task set please refer to the [[2008:Audio Tag Classification]] page.  
  
 
===General Legend===
 
===General Legend===
 
====Team ID====
 
====Team ID====
'''LB''' = [https://www.music-ir.org/mirex/2008/abs/.pdf L. Barrington, D. Turnball, G. Lanckriet]<br />
+
'''LB''' = [https://www.music-ir.org/mirex/abstracts/2008/AT_barrington.pdf L. Barrington, D. Turnbull, G. Lanckriet]<br />
'''BBE 1''' = [https://www.music-ir.org/mirex/2008/abs/.pdf T. Bertin-Mahieux, Y. Bengio, D. Eck (KNN)]<br />
+
'''BBE 1''' = [https://www.music-ir.org/mirex/abstracts/2008/mirex_knn.pdf T. Bertin-Mahieux, Y. Bengio, D. Eck (KNN)]<br />
'''BBE 2''' = [https://www.music-ir.org/mirex/2008/abs/.pdf T. Bertin-Mahieux, Y. Bengio, D. Eck (NNet)]<br />
+
'''BBE 2''' = [https://www.music-ir.org/mirex/abstracts/2008/mirex_nnet.pdf T. Bertin-Mahieux, Y. Bengio, D. Eck (NNet)]<br />
'''BBE 3''' = [https://www.music-ir.org/mirex/2008/abs/.pdf T. Bertin-Mahieux, D. Eck, P. Lamere, Y. Bengio (Thierry/Lamere Boosting)]<br />
+
'''BBE 3''' = [https://www.music-ir.org/mirex/abstracts/2008/mirex_boosters.pdf T. Bertin-Mahieux, D. Eck, P. Lamere, Y. Bengio (Thierry/Lamere Boosting)]<br />
'''TB''' = [https://www.music-ir.org/mirex/2008/abs/.pdf T. Bertin-Mahieux (dumb/smurf)]<br />
+
'''TB''' = [https://www.music-ir.org/mirex/abstracts/2008/mirex_smurfs.pdf T. Bertin-Mahieux (dumb/smurf)]<br />
'''ME1''' = [https://www.music-ir.org/mirex/2008/abs/AI_CC_GC_MC_AS_mandel.pdf M. I. Mandel, D. P. W. Ellis 1]<br />
+
'''ME1''' = [https://www.music-ir.org/mirex/abstracts/2008/AA_AG_AT_MM_CC_mandel.pdf M. I. Mandel, D. P. W. Ellis 1]<br />
'''ME2''' = [https://www.music-ir.org/mirex/2008/abs/AI_CC_GC_MC_AS_mandel.pdf M. I. Mandel, D. P. W. Ellis 2]<br />
+
'''ME2''' = [https://www.music-ir.org/mirex/abstracts/2008/AA_AG_AT_MM_CC_mandel.pdf M. I. Mandel, D. P. W. Ellis 2]<br />
'''ME3''' = [https://www.music-ir.org/mirex/2008/abs/AI_CC_GC_MC_AS_mandel.pdf M. I. Mandel, D. P. W. Ellis 3]<br />
+
'''ME3''' = [https://www.music-ir.org/mirex/abstracts/2008/AA_AG_AT_MM_CC_mandel.pdf M. I. Mandel, D. P. W. Ellis 3]<br />
'''GP1''' = [https://www.music-ir.org/mirex/2008/abs/.pdf G. Peeters 1]<br />
+
'''GP1''' = [https://www.music-ir.org/mirex/abstracts/2008/Peeters_2008_ISMIR_MIREX.pdf G. Peeters 1]<br />
'''GP2''' = [https://www.music-ir.org/mirex/2008/abs/.pdf G. Peeters 2]<br />
+
'''GP2''' = [https://www.music-ir.org/mirex/abstracts/2008/Peeters_2008_ISMIR_MIREX.pdf G. Peeters 2]<br />
'''TTKV''' = [https://www.music-ir.org/mirex/2008/abs/.pdf K. Trohidis, G. Tsoumakas, G. Kalliris, I. Vlahavas]<br />
+
'''TTKV''' = [https://www.music-ir.org/mirex/abstracts/2008/auth.pdf K. Trohidis, G. Tsoumakas, G. Kalliris, I. Vlahavas]<br />
  
 
==Overall Summary Results==
 
==Overall Summary Results==
  
<csv>tag/tag.grand.summary.show.csv</csv>
+
<csv>2008/tag/tag.grand.summary.show.csv</csv>
  
  
 
===Summary Positive Example Accuracy (Average Across All Folds)===
 
===Summary Positive Example Accuracy (Average Across All Folds)===
  
<csv>tag/rounded/tag.binary_avg_positive_example_Accuracy.csv</csv>
+
<csv>2008/tag/rounded/tag.binary_avg_positive_example_Accuracy.csv</csv>
  
 
===Summary Negative Example Accuracy (Average Across All Folds)===
 
===Summary Negative Example Accuracy (Average Across All Folds)===
  
<csv>tag/rounded/tag.binary_avg_negative_example_Accuracy.csv</csv>
+
<csv>2008/tag/rounded/tag.binary_avg_negative_example_Accuracy.csv</csv>
  
 
===Summary Binary relevance F-Measure (Average Across All Folds)===
 
===Summary Binary relevance F-Measure (Average Across All Folds)===
  
<csv>tag/rounded/tag.binary_avg_Fmeasure.csv</csv>
+
<csv>2008/tag/rounded/tag.binary_avg_Fmeasure.csv</csv>
  
 
===Summary Binary Accuracy (Average Across All Folds)===
 
===Summary Binary Accuracy (Average Across All Folds)===
  
<csv>tag/rounded/tag.binary_avg_Accuracy.csv</csv>
+
<csv>2008/tag/rounded/tag.binary_avg_Accuracy.csv</csv>
  
 
===Summary AUC-ROC Tag (Average Across All Folds)===
 
===Summary AUC-ROC Tag (Average Across All Folds)===
  
<csv>tag/rounded/tag.affinity_tag_AUC_ROC.csv</csv>
+
<csv>2008/tag/rounded/tag.affinity_tag_AUC_ROC.csv</csv>
  
 
==Friedman test results==
 
==Friedman test results==
Line 46: Line 46:
 
The following table and plot show the results of Friedman's ANOVA with Tukey-Kramer multiple comparisons computed over the Area Under the ROC curve (AUC-ROC) for each '''tag''' in the test, averaged over all folds.
 
The following table and plot show the results of Friedman's ANOVA with Tukey-Kramer multiple comparisons computed over the Area Under the ROC curve (AUC-ROC) for each '''tag''' in the test, averaged over all folds.
  
<csv>tag/friedmansTables/tag.affinity.AUC_ROC_TAG.friedman.tukeyKramerHSD.csv</csv>
+
<csv>2008/tag/friedmansTables/tag.affinity.AUC_ROC_TAG.friedman.tukeyKramerHSD.csv</csv>
  
[[Image:Affinity.AUC_ROC_TAG.friedman.tukeyKramerHSD.png]]
+
[[Image:2008_affinity.auc_roc_tag.friedman.tukeykramerhsd.png]]
  
 
===AUC-ROC Track Friedman test===
 
===AUC-ROC Track Friedman test===
 
The following table and plot show the results of Friedman's ANOVA with Tukey-Kramer multiple comparisons computed over the Area Under the ROC curve (AUC-ROC) for each '''track''' in the test. Each track appears in exactly once over all three folds of the test. However, we are uncertain if these measurements are truly independent as an multiple tracks from each artist are used.
 
The following table and plot show the results of Friedman's ANOVA with Tukey-Kramer multiple comparisons computed over the Area Under the ROC curve (AUC-ROC) for each '''track''' in the test. Each track appears in exactly once over all three folds of the test. However, we are uncertain if these measurements are truly independent as an multiple tracks from each artist are used.
  
<csv>tag/friedmansTables/tag.affinity.AUC_ROC_TRACK.friedman.tukeyKramerHSD.csv</csv>
+
<csv>2008/tag/friedmansTables/tag.affinity.AUC_ROC_TRACK.friedman.tukeyKramerHSD.csv</csv>
  
[[Image:Affinity.AUC_ROC_TRACK.friedman.tukeyKramerHSD.png]]
+
[[Image:2008_affinity.auc_roc_track.friedman.tukeykramerhsd.png]]
  
 
===Tag Classification Accuracy Friedman test===
 
===Tag Classification Accuracy Friedman test===
 
The following table and plot show the results of Friedman's ANOVA with Tukey-Kramer multiple comparisons computed over the classification accuracy for each '''tag''' in the test, averaged over all folds.
 
The following table and plot show the results of Friedman's ANOVA with Tukey-Kramer multiple comparisons computed over the classification accuracy for each '''tag''' in the test, averaged over all folds.
  
<csv>tag/friedmansTables/tag.binary_Accuracy.friedman.tukeyKramerHSD.csv</csv>
+
<csv>2008/tag/friedmansTables/tag.binary_Accuracy.friedman.tukeyKramerHSD.csv</csv>
  
[[Image:Binary_Accuracy.friedman.tukeyKramerHSD.png]]
+
[[Image:2008_binary_accuracy.friedman.tukeykramerhsd.png]]
  
 
===Tag F-measure Friedman test===
 
===Tag F-measure Friedman test===
 
The following table and plot show the results of Friedman's ANOVA with Tukey-Kramer multiple comparisons computed over the F-measure for each '''tag''' in the test, averaged over all folds.
 
The following table and plot show the results of Friedman's ANOVA with Tukey-Kramer multiple comparisons computed over the F-measure for each '''tag''' in the test, averaged over all folds.
  
<csv>tag/friedmansTables/tag.binary_FMeasure.friedman.tukeyKramerHSD.csv</csv>
+
<csv>2008/tag/friedmansTables/tag.binary_FMeasure.friedman.tukeyKramerHSD.csv</csv>
  
[[Image:Binary_FMeasure.friedman.tukeyKramerHSD.png]]
+
[[Image:2008_binary_fmeasure.friedman.tukeykramerhsd.png]]
 +
 
 +
 
 +
==Beta-Binomial test results==
 +
 
 +
===Accuracy on positive examples Beta-Binomial results===
 +
The following table and plot show the results of simulations from the Beta-Binomial model using the accuracy of each algorithm's classification only on the positive examples.  It only shows the relative proportion of true positives and false negatives, and should be considered with the classification accuracy on the negative examples.  The image shows the estimate of the overall performance with 95% confidence intervals.
 +
 
 +
 
 +
<csv>2008/tag/tag.binary.per.fold.positive.example.accuracy.betabinomial.csv</csv>
 +
 
 +
 
 +
[[Image:binary_per_fold_positive_example_Accuracy.png]]
 +
 
 +
 
 +
The plots for each tag are more interesting and the 95% confidence intervals are much tighter.  Since there are so many of them, it is difficult to post them to the wiki.  You can download a tar.gz zip file containing all of them [https://www.music-ir.org/mirex/results/2008/tag/binary_positive_example_Accuracy.betaBinomial.images.tar.gz here].
 +
 
 +
===Accuracy on negative examples Beta-Binomial results===
 +
The following table and plot show the results of simulations from the Beta-Binomial model using the accuracy of each algorithm's classification only on the negative examples.  It only shows the relative proportion of true negatives and false positives, and should be considered with the classification accuracy on the positive examples.  The image shows the estimate of the overall performance with 95% confidence intervals.
 +
 
 +
 
 +
<csv>2008/tag/tag.binary.per.fold.negative.example.accuracy.betabinomial.csv</csv>
 +
 
 +
 
 +
[[Image:binary_per_fold_negative_example_Accuracy.png]]
 +
 
 +
 
 +
The plots for each tag are more interesting and the 95% confidence intervals are much tighter.  Since there are so many of them, it is difficult to post them to the wiki.  You can download a tar.gz file containing all of them [https://www.music-ir.org/mirex/results/2008/tag/binary_negative_example_Accuracy.betaBinomial.images.tar.gz here].
  
 
==Assorted Results Files for Download==
 
==Assorted Results Files for Download==
 
===AUC-ROC Clip Data===
 
===AUC-ROC Clip Data===
 
(Too large for easy Wiki viewing)<br \>
 
(Too large for easy Wiki viewing)<br \>
[https://www.music-ir.org/mirex/2008/results/tag/rounded/tag.affinity_clip_AUC_ROC.csv tag.affinity_clip_AUC_ROC.csv]<br />
+
[https://www.music-ir.org/mirex/results/2008/tag/rounded/tag.affinity_clip_AUC_ROC.csv tag.affinity_clip_AUC_ROC.csv]<br />
  
 
===CSV Files Without Rounding (Averaged across folds)===
 
===CSV Files Without Rounding (Averaged across folds)===
[https://www.music-ir.org/mirex/2008/results/tag/csv_raw/tag.affinity.tag.auc.roc.csv tag.affinity.clip.auc.roc.csv]<br />
+
[https://www.music-ir.org/mirex/results/2008/tag/csv_raw/tag.affinity.tag.auc.roc.csv tag.affinity.clip.auc.roc.csv]<br />
[https://www.music-ir.org/mirex/2008/results/tag/csv_raw/tag.binary.avg.accuracy.csv tag.binary.accuracy.csv]<br />
+
[https://www.music-ir.org/mirex/results/2008/tag/csv_raw/tag.affinity.clip.auc.roc.csv tag.affinity.clip.auc.roc.csv]<br />
[https://www.music-ir.org/mirex/2008/results/tag/csv_raw/tag.binary.avg.fmeasure.csv tag.binary.fmeasure.csv]<br />
+
[https://www.music-ir.org/mirex/results/2008/tag/csv_raw/tag.binary.avg.accuracy.csv tag.binary.accuracy.csv]<br />
[https://www.music-ir.org/mirex/2008/results/tag/csv_raw/tag.binary.avg.negative.example.accuracy.csv tag.binary.negative.example.accuracy.csv]<br />
+
[https://www.music-ir.org/mirex/results/2008/tag/csv_raw/tag.binary.avg.fmeasure.csv tag.binary.fmeasure.csv]<br />
[https://www.music-ir.org/mirex/2008/results/tag/csv_raw/tag.binary.avg.positive.example.accuracy.csv tag.binary.positive.example.accuracy.csv]<br />
+
[https://www.music-ir.org/mirex/results/2008/tag/csv_raw/tag.binary.avg.negative.example.accuracy.csv tag.binary.negative.example.accuracy.csv]<br />
 +
[https://www.music-ir.org/mirex/results/2008/tag/csv_raw/tag.binary.avg.positive.example.accuracy.csv tag.binary.positive.example.accuracy.csv]<br />
 +
 
 +
===CSV Files Without Rounding (Fold information)===
 +
[https://www.music-ir.org/mirex/results/2008/tag/csv_raw/tag.binary.per.fold.positive.example.accuracy.csv tag.binary.per.fold.positive.example.accuracy.csv]<br />
 +
[https://www.music-ir.org/mirex/results/2008/tag/csv_raw/tag.binary.per.fold.negative.example.accuracy.csv tag.binary.per.fold.negative.example.accuracy.csv]<br />
 +
[https://www.music-ir.org/mirex/results/2008/tag/csv_raw/tag.binary.per.fold.fmeasure.csv tag.binary.per.fold.fmeasure.csv]<br />
 +
[https://www.music-ir.org/mirex/results/2008/tag/csv_raw/tag.binary.per.fold.accuracy.csv tag.binary.per.fold.accuracy.csv]<br />
 +
[https://www.music-ir.org/mirex/results/2008/tag/csv_raw/tag.affinity.tag.per.fold.auc.roc.csv tag.affinity.tag.per.fold.auc.roc.csv]<br />
  
 
===Results By Algorithm===
 
===Results By Algorithm===
 
(.tar.gz) <br />
 
(.tar.gz) <br />
'''LB''' = [https://www.music-ir.org/mirex/2008/results/tag/detailedReports/LB.tar.gz L. Barrington, D. Turnball, G. Lanckriet]<br />
+
'''LB''' = [https://www.music-ir.org/mirex/results/2008/tag/detailedReports/LB.tar.gz L. Barrington, D. Turnbull, G. Lanckriet]<br />
'''BBE 1''' = [https://www.music-ir.org/mirex/2008/results/tag/detailedReports/BBE1.tar.gz T. Bertin-Mahieux, Y. Bengio, D. Eck (KNN)]<br />
+
'''BBE 1''' = [https://www.music-ir.org/mirex/results/2008/tag/detailedReports/BBE1.tar.gz T. Bertin-Mahieux, Y. Bengio, D. Eck (KNN)]<br />
'''BBE 2''' = [https://www.music-ir.org/mirex/2008/results/tag/detailedReports/BBE2.tar.gz T. Bertin-Mahieux, Y. Bengio, D. Eck (NNet)]<br />
+
'''BBE 2''' = [https://www.music-ir.org/mirex/results/2008/tag/detailedReports/BBE2.tar.gz T. Bertin-Mahieux, Y. Bengio, D. Eck (NNet)]<br />
'''BBE 3''' = [https://www.music-ir.org/mirex/2008/results/tag/detailedReports/BBE3.tar.gz T. Bertin-Mahieux, D. Eck, P. Lamere, Y. Bengio (Thierry/Lamere Boosting)]<br />
+
'''BBE 3''' = [https://www.music-ir.org/mirex/results/2008/tag/detailedReports/BBE3.tar.gz T. Bertin-Mahieux, D. Eck, P. Lamere, Y. Bengio (Thierry/Lamere Boosting)]<br />
'''TB''' = [https://www.music-ir.org/mirex/2008/results/tag/detailedReports/TB.tar.gz Bertin-Mahieux (dumb/smurf)]<br />
+
'''TB''' = [https://www.music-ir.org/mirex/results/2008/tag/detailedReports/TB.tar.gz Bertin-Mahieux (dumb/smurf)]<br />
'''ME1''' = [https://www.music-ir.org/mirex/2008/results/tag/detailedReports/ME1.tar.gz M. I. Mandel, D. P. W. Ellis 1]<br />
+
'''ME1''' = [https://www.music-ir.org/mirex/results/2008/tag/detailedReports/ME1.tar.gz M. I. Mandel, D. P. W. Ellis 1]<br />
'''ME2''' = [https://www.music-ir.org/mirex/2008/results/tag/detailedReports/ME2.tar.gz M. I. Mandel, D. P. W. Ellis 2]<br />
+
'''ME2''' = [https://www.music-ir.org/mirex/results/2008/tag/detailedReports/ME2.tar.gz M. I. Mandel, D. P. W. Ellis 2]<br />
'''ME3''' = [https://www.music-ir.org/mirex/2008/results/tag/detailedReports/ME3.tar.gz M. I. Mandel, D. P. W. Ellis 3]<br />
+
'''ME3''' = [https://www.music-ir.org/mirex/results/2008/tag/detailedReports/ME3.tar.gz M. I. Mandel, D. P. W. Ellis 3]<br />
'''GP1''' = [https://www.music-ir.org/mirex/2008/results/tag/detailedReports/GP1.tar.gz G. Peeters 1]<br />
+
'''GP1''' = [https://www.music-ir.org/mirex/results/2008/tag/detailedReports/GP1.tar.gz G. Peeters 1]<br />
'''GP2''' = [https://www.music-ir.org/mirex/2008/results/tag/detailedReports/GP2.tar.gz G. Peeters 2]<br />
+
'''GP2''' = [https://www.music-ir.org/mirex/results/2008/tag/detailedReports/GP2.tar.gz G. Peeters 2]<br />
'''TTKV''' = [https://www.music-ir.org/mirex/2008/results/tag/detailedReports/TTKV.tar.gz K. Trohidis, G. Tsoumakas, G. Kalliris, I. Vlahavas]<br />
+
'''TTKV''' = [https://www.music-ir.org/mirex/results/2008/tag/detailedReports/TTKV.tar.gz K. Trohidis, G. Tsoumakas, G. Kalliris, I. Vlahavas]<br />
  
  

Latest revision as of 13:54, 7 June 2010

Introduction

These are the results for the 2008 running of the Audio Tag Classification task. For background information about this task set please refer to the 2008:Audio Tag Classification page.

General Legend

Team ID

LB = L. Barrington, D. Turnbull, G. Lanckriet
BBE 1 = T. Bertin-Mahieux, Y. Bengio, D. Eck (KNN)
BBE 2 = T. Bertin-Mahieux, Y. Bengio, D. Eck (NNet)
BBE 3 = T. Bertin-Mahieux, D. Eck, P. Lamere, Y. Bengio (Thierry/Lamere Boosting)
TB = T. Bertin-Mahieux (dumb/smurf)
ME1 = M. I. Mandel, D. P. W. Ellis 1
ME2 = M. I. Mandel, D. P. W. Ellis 2
ME3 = M. I. Mandel, D. P. W. Ellis 3
GP1 = G. Peeters 1
GP2 = G. Peeters 2
TTKV = K. Trohidis, G. Tsoumakas, G. Kalliris, I. Vlahavas

Overall Summary Results

Measure BBE1 BBE2 BBE3 ME1 ME2 ME3 GP1 GP2 LB TB TTKV
Average Tag Positive Example Accuracy 0.05 1.00 0.85 0.67 0.68 0.66 0.04 0.03 0.28 0.91 0.03
Average Tag Negative Example Accuracy 0.99 0.00 0.37 0.71 0.73 0.73 0.98 0.98 0.94 0.09 0.97
Average Tag F-Measure 0.06 0.15 0.19 0.24 0.26 0.26 0.03 0.02 0.28 0.15 0.04
Average Tag Accuracy 0.91 0.09 0.43 0.71 0.73 0.72 0.90 0.89 0.90 0.17 0.90
Average AUC-ROC Clip 0.82 0.49 0.81 0.77 0.79 0.78 n/a n/a 0.84 0.69 0.78
Average AUC-ROC Tag 0.66 0.50 0.74 0.75 0.77 0.76 n/a n/a 0.77 0.50 0.50
Overall Beta-Binomial Positive Example Accuracy 0.01 1.00 0.89 0.68 0.69 0.67 0.00 0.00 0.26 1.00 0.00
Overall Beta-Binomial Negative Example Accuracy 1.00 0.00 0.35 0.72 0.74 0.73 1.00 1.00 0.95 0.00 1.00

download these results as csv


Summary Positive Example Accuracy (Average Across All Folds)

Tag Positive examples Negative examples BBE1 BBE2 BBE3 ME1 ME2 ME3 GP1 GP2 LB TB TTKV
metal 35 2180 0 1 0.75 0.78 0.66 0.67 0 0.1 0.18 0.33 0
ambient 60 2155 0 0.99 0.43 0.56 0.62 0.64 0 0.08 0.27 1 0
fast 109 2106 0 1 0.89 0.7 0.7 0.74 0 0.02 0.22 1 0
solo 53 2162 0 1 0.98 0.61 0.59 0.59 0 0 0.1 1 0
jazz 138 2077 0.05 1 0.91 0.7 0.74 0.72 0.04 0 0.4 1 0.02
instrumental 102 2113 0 1 0.73 0.59 0.62 0.59 0.33 0.04 0.08 1 0
horns 40 2175 0 1 0.8 0.54 0.54 0.47 0 0.06 0.02 0.67 0
house 65 2150 0 1 0.87 0.82 0.77 0.74 0 0.06 0.2 1 0
male 738 1477 0.12 1 1 0.71 0.72 0.72 0.01 0 0.49 1 0.2
beat 137 2078 0 1 0.8 0.74 0.78 0.74 0 0.04 0.25 1 0
strings 55 2160 0.08 1 0.85 0.71 0.73 0.78 0 0 0.22 1 0
saxophone 68 2147 0 1 0.85 0.8 0.82 0.72 0 0 0.33 1 0
piano 185 2030 0.06 1 0.65 0.67 0.72 0.74 0 0 0.34 1 0
loud 37 2178 0 1 0.8 0.71 0.67 0.71 0 0 0.17 0.67 0
noise 42 2173 0.26 0.97 0.71 0.59 0.51 0.55 0 0 0.16 0.67 0
80s 117 2098 0.01 1 0.93 0.6 0.59 0.63 0 0.17 0.23 1 0
pop 479 1736 0.1 1 0.98 0.67 0.69 0.71 0.02 0 0.45 1 0
female 345 1870 0.12 1 0.95 0.75 0.72 0.73 0.01 0.02 0.44 1 0
slow 159 2056 0 1 0.97 0.64 0.69 0.68 0.08 0 0.33 1 0
funk 37 2178 0 1 0.74 0.72 0.75 0.69 0 0.02 0.03 0.67 0
keyboard 42 2173 0 1 0.91 0.45 0.54 0.46 0 0 0 1 0
electronica 166 2049 0.06 1 0.94 0.68 0.72 0.68 0 0.03 0.28 1 0
rock 664 1551 0.24 1 0.98 0.75 0.75 0.73 0.05 0.01 0.69 1 0.24
vocal 270 1945 0 1 0.94 0.61 0.63 0.6 0 0 0.22 1 0
acoustic 38 2177 0 1 0.74 0.65 0.62 0.64 0 0 0.18 0.33 0
hip hop 152 2063 0.12 1 0.9 0.83 0.86 0.82 0.23 0 0.58 1 0.04
guitar 849 1366 0.12 1 0.99 0.72 0.72 0.73 0.15 0 0.64 1 0.37
bass 420 1795 0 1 0.95 0.5 0.53 0.54 0 0.07 0.23 1 0
british 82 2133 0.01 1 0.84 0.62 0.63 0.66 0 0 0.11 1 0
dance 324 1891 0.04 1 0.99 0.73 0.75 0.72 0.04 0 0.54 1 0.04
r&b 38 2177 0.02 1 0.56 0.6 0.65 0.7 0.02 0 0.18 0.33 0
electronic 485 1730 0.2 1 0.97 0.68 0.7 0.69 0.18 0 0.52 1 0.16
drum 927 1288 0.01 1 1 0.61 0.59 0.57 0.35 0 0.5 1 0.28
soft 62 2153 0 1 0.92 0.74 0.81 0.69 0.09 0 0.2 1 0
punk 51 2164 0 1 0.71 0.76 0.71 0.63 0 0 0.3 1 0
country 74 2141 0.18 1 0.98 0.6 0.62 0.61 0.05 0 0.15 1 0
drum machine 89 2126 0 1 0.89 0.66 0.69 0.69 0 0 0.21 1 0
voice 148 2067 0 1 0.97 0.53 0.59 0.52 0 0.36 0.08 1 0
quiet 43 2172 0.16 0.86 0.55 0.63 0.75 0.7 0 0.13 0.25 1 0
distortion 60 2155 0 1 0.82 0.68 0.63 0.64 0 0 0.13 1 0
synth 480 1735 0.02 1 0.97 0.68 0.63 0.62 0 0 0.4 1 0.02
rap 157 2058 0.46 1 0.68 0.87 0.88 0.85 0 0 0.65 1 0.05
techno 248 1967 0 1 0.92 0.76 0.8 0.75 0 0 0.5 1 0.02
organ 35 2180 0 1 0.68 0.4 0.31 0.33 0 0 0 0.67 0
trumpet 39 2176 0 1 0.81 0.63 0.69 0.58 0 0 0.14 0.67 0
MEAN 198.76 2016.24 0.05 1 0.85 0.67 0.68 0.66 0.04 0.03 0.28 0.91 0.03

download these results as csv

Summary Negative Example Accuracy (Average Across All Folds)

Tag Positive examples Negative examples BBE1 BBE2 BBE3 ME1 ME2 ME3 GP1 GP2 LB TB TTKV
metal 35 2180 1 0 0.64 0.79 0.79 0.79 1 0.97 0.99 0.67 1
ambient 60 2155 1 0 0.87 0.75 0.81 0.75 1 0.97 0.98 0 1
fast 109 2106 1 0 0.41 0.68 0.69 0.7 1 0.97 0.96 0 1
solo 53 2162 1 0 0.17 0.74 0.73 0.74 1 1 0.98 0 1
jazz 138 2077 0.99 0 0.47 0.85 0.87 0.87 1 1 0.96 0 0.98
instrumental 102 2113 1 0 0.37 0.63 0.65 0.64 0.67 0.95 0.96 0 1
horns 40 2175 1 0 0.31 0.66 0.67 0.64 1 0.94 0.98 0.33 1
house 65 2150 1 0 0.41 0.76 0.8 0.76 1 0.96 0.98 0 1
male 738 1477 0.91 0 0.05 0.72 0.74 0.74 0.99 1 0.75 0 0.8
beat 137 2078 1 0 0.55 0.7 0.74 0.75 1 0.95 0.95 0 1
strings 55 2160 0.97 0 0.46 0.76 0.75 0.77 1 1 0.98 0 1
saxophone 68 2147 1 0 0.45 0.85 0.85 0.84 1 0.99 0.98 0 1
piano 185 2030 0.96 0 0.44 0.65 0.68 0.69 1 1 0.94 0 1
loud 37 2178 1 0 0.56 0.81 0.82 0.81 1 1 0.99 0.33 1
noise 42 2173 0.95 0 0.69 0.76 0.78 0.78 1 0.99 0.98 0.33 1
80s 117 2098 1 0 0.13 0.68 0.68 0.7 1 0.92 0.96 0 1
pop 479 1736 0.93 0 0.1 0.68 0.68 0.68 0.99 1 0.85 0 1
female 345 1870 0.93 0 0.17 0.81 0.82 0.8 1 0.98 0.9 0 1
slow 159 2056 1 0 0.27 0.65 0.68 0.68 0.98 1 0.95 0 1
funk 37 2178 1 0 0.43 0.71 0.71 0.7 1 0.98 0.98 0.33 1
keyboard 42 2173 1 0 0.16 0.59 0.56 0.53 1 1 0.98 0 1
electronica 166 2049 0.99 0 0.39 0.68 0.73 0.73 1 0.96 0.94 0 1
rock 664 1551 0.92 0 0.28 0.79 0.81 0.8 0.99 0.99 0.87 0 0.72
vocal 270 1945 1 0 0.08 0.58 0.61 0.6 1 0.99 0.89 0 1
acoustic 38 2177 1 0 0.45 0.68 0.68 0.66 1 0.96 0.99 0.67 1
hip hop 152 2063 0.99 0 0.67 0.83 0.87 0.87 0.99 1 0.97 0 0.95
guitar 849 1366 0.94 0.01 0.22 0.71 0.73 0.73 0.95 1 0.78 0 0.62
bass 420 1795 1 0 0.05 0.57 0.58 0.6 1 0.94 0.82 0 1
british 82 2133 1 0 0.22 0.69 0.7 0.67 1 1 0.97 0 1
dance 324 1891 1 0 0.19 0.75 0.8 0.8 0.99 1 0.92 0 0.94
r&b 38 2177 0.99 0 0.66 0.75 0.76 0.74 1 1 0.99 0.67 1
electronic 485 1730 0.97 0 0.27 0.72 0.77 0.77 0.96 1 0.86 0 0.83
drum 927 1288 1 0.01 0.02 0.56 0.57 0.6 0.66 1 0.64 0 0.7
soft 62 2153 1 0 0.3 0.68 0.69 0.68 0.99 1 0.98 0 1
punk 51 2164 1 0 0.6 0.8 0.82 0.81 1 1 0.98 0 1
country 74 2141 0.95 0 0.19 0.73 0.73 0.76 0.98 1 0.97 0 1
drum machine 89 2126 1 0 0.43 0.65 0.69 0.68 1 1 0.97 0 1
voice 148 2067 1 0 0.06 0.52 0.54 0.54 1 0.65 0.93 0 1
quiet 43 2172 0.99 4.42E-004 0.83 0.84 0.85 0.85 1 0.97 0.99 0 1
distortion 60 2155 1 0 0.53 0.77 0.78 0.78 1 1 0.98 0 1
synth 480 1735 1 0 0.14 0.65 0.68 0.68 1 1 0.83 0 0.98
rap 157 2058 0.97 0 0.64 0.85 0.89 0.89 1 1 0.97 0 0.95
techno 248 1967 1 0 0.52 0.74 0.8 0.8 1 1 0.94 0 0.97
organ 35 2180 1 0 0.39 0.55 0.55 0.53 1 0.99 0.98 0.33 1
trumpet 39 2176 1 0 0.39 0.77 0.76 0.74 1 1 0.98 0.33 1
MEAN 198.76 2016.24 0.99 0 0.37 0.71 0.73 0.73 0.98 0.98 0.94 0.09 0.97

download these results as csv

Summary Binary relevance F-Measure (Average Across All Folds)

Tag Positive Examples Negative Examples BBE1 BBE2 BBE3 ME1 ME2 ME3 GP1 GP2 LB TB TTKV
metal 35 2180 0 0.03 0.09 0.1 0.09 0.09 0 0.06 0.17 0.01 0
ambient 60 2155 0 0.05 0.1 0.09 0.14 0.12 0 0.07 0.26 0.05 0
fast 109 2106 0 0.09 0.14 0.17 0.18 0.19 0 0.02 0.21 0.09 0
solo 53 2162 0 0.05 0.06 0.1 0.09 0.1 0 0 0.1 0.05 0
jazz 138 2077 0.08 0.12 0.21 0.36 0.4 0.39 0.07 0 0.4 0.12 0.02
instrumental 102 2113 0 0.09 0.1 0.13 0.14 0.13 0.03 0.02 0.08 0.09 0
horns 40 2175 0 0.04 0.04 0.05 0.06 0.05 0 0.02 0.02 0.02 0
house 65 2150 0 0.06 0.08 0.17 0.19 0.15 0 0.04 0.2 0.06 0
male 738 1477 0.17 0.5 0.51 0.63 0.64 0.65 0.02 0 0.49 0.5 0.24
beat 137 2078 0 0.12 0.2 0.24 0.28 0.27 0 0.03 0.25 0.12 0
strings 55 2160 0.05 0.05 0.08 0.13 0.13 0.14 0 0 0.22 0.05 0
saxophone 68 2147 0 0.06 0.09 0.26 0.26 0.22 0 0 0.31 0.06 0
piano 185 2030 0.07 0.15 0.11 0.24 0.27 0.29 0 0 0.34 0.15 0
loud 37 2178 0 0.03 0.1 0.11 0.11 0.11 0 0 0.19 0.02 0
noise 42 2173 0.05 0.04 0.09 0.08 0.08 0.09 0 0 0.15 0.02 0
80s 117 2098 0.01 0.1 0.11 0.16 0.16 0.18 0 0.05 0.23 0.1 0
pop 479 1736 0.12 0.36 0.38 0.47 0.49 0.5 0.04 0 0.45 0.36 0
female 345 1870 0.14 0.27 0.29 0.54 0.52 0.51 0.01 0.03 0.42 0.27 0.01
slow 159 2056 0 0.13 0.17 0.21 0.24 0.23 0.08 0 0.33 0.13 0
funk 37 2178 0 0.03 0.05 0.08 0.08 0.07 0 0.02 0.03 0.02 0
keyboard 42 2173 0 0.04 0.04 0.04 0.05 0.04 0 0 0 0.04 0
electronica 166 2049 0.09 0.14 0.2 0.24 0.28 0.27 0 0.03 0.27 0.14 0
rock 664 1551 0.27 0.46 0.54 0.67 0.68 0.67 0.08 0.02 0.69 0.46 0.26
vocal 270 1945 0 0.22 0.22 0.26 0.28 0.27 0 0 0.22 0.22 0
acoustic 38 2177 0 0.03 0.05 0.07 0.06 0.06 0 0 0.18 0.01 0
hip hop 152 2063 0.15 0.13 0.28 0.4 0.47 0.46 0.22 0 0.56 0.13 0.05
guitar 849 1366 0.18 0.56 0.61 0.66 0.67 0.67 0.18 0 0.64 0.55 0.37
bass 420 1795 0 0.32 0.32 0.3 0.32 0.33 0 0.1 0.23 0.32 0
british 82 2133 0.02 0.07 0.08 0.13 0.13 0.13 0 0 0.1 0.07 0
dance 324 1891 0.07 0.25 0.29 0.46 0.52 0.5 0.06 0 0.54 0.25 0.06
r&b 38 2177 0.04 0.03 0.06 0.08 0.09 0.09 0.04 0 0.18 0.01 0
electronic 485 1730 0.28 0.36 0.42 0.5 0.55 0.55 0.19 0 0.51 0.36 0.18
drum 927 1288 0.01 0.59 0.6 0.55 0.54 0.53 0.23 0 0.5 0.59 0.33
soft 62 2153 0 0.05 0.07 0.11 0.13 0.11 0.08 0 0.2 0.05 0
punk 51 2164 0 0.05 0.08 0.15 0.15 0.13 0 0 0.29 0.05 0
country 74 2141 0.13 0.06 0.08 0.14 0.14 0.15 0.03 0 0.15 0.06 0
drum machine 89 2126 0 0.08 0.12 0.13 0.15 0.15 0 0 0.22 0.08 0
voice 148 2067 0 0.13 0.13 0.13 0.15 0.13 0 0.12 0.08 0.13 0
quiet 43 2172 0.19 0.03 0.21 0.13 0.16 0.15 0 0.09 0.25 0.04 0
distortion 60 2155 0 0.05 0.09 0.14 0.13 0.13 0 0 0.12 0.05 0
synth 480 1735 0.03 0.36 0.38 0.46 0.45 0.45 0 0 0.39 0.35 0.04
rap 157 2058 0.48 0.13 0.21 0.46 0.52 0.52 0 0 0.63 0.13 0.06
techno 248 1967 0 0.2 0.33 0.39 0.47 0.45 0 0 0.49 0.2 0.02
organ 35 2180 0 0.03 0.03 0.03 0.02 0.02 0 0 0 0.02 0
trumpet 39 2176 0 0.03 0.05 0.08 0.09 0.07 0 0 0.13 0.02 0
MEAN 198.76 2016.24 0.06 0.15 0.19 0.24 0.26 0.26 0.03 0.02 0.28 0.15 0.04

download these results as csv

Summary Binary Accuracy (Average Across All Folds)

Tag Positive Examples Negative Examples BBE1 BBE2 BBE3 ME1 ME2 ME3 GP1 GP2 LB TB TTKV
metal 35 2180 0.98 0.02 0.64 0.79 0.79 0.78 0.98 0.95 0.97 0.66 0.98
ambient 60 2155 0.97 0.03 0.86 0.74 0.8 0.74 0.97 0.95 0.96 0.03 0.97
fast 109 2106 0.95 0.05 0.43 0.68 0.69 0.7 0.95 0.92 0.92 0.05 0.95
solo 53 2162 0.98 0.03 0.19 0.74 0.72 0.73 0.98 0.98 0.96 0.02 0.98
jazz 138 2077 0.93 0.07 0.5 0.84 0.86 0.86 0.94 0.94 0.93 0.06 0.92
instrumental 102 2113 0.95 0.05 0.39 0.63 0.65 0.63 0.65 0.91 0.92 0.05 0.95
horns 40 2175 0.98 0.02 0.32 0.66 0.67 0.64 0.98 0.92 0.96 0.34 0.98
house 65 2150 0.97 0.03 0.43 0.76 0.8 0.76 0.97 0.93 0.95 0.03 0.97
male 738 1477 0.64 0.34 0.37 0.72 0.73 0.74 0.66 0.66 0.66 0.34 0.59
beat 137 2078 0.94 0.06 0.57 0.71 0.75 0.75 0.94 0.9 0.91 0.06 0.94
strings 55 2160 0.95 0.03 0.47 0.76 0.75 0.77 0.97 0.98 0.96 0.02 0.98
saxophone 68 2147 0.97 0.03 0.46 0.85 0.85 0.84 0.97 0.95 0.96 0.03 0.97
piano 185 2030 0.88 0.09 0.45 0.66 0.68 0.7 0.92 0.92 0.89 0.08 0.92
loud 37 2178 0.98 0.02 0.56 0.81 0.82 0.81 0.98 0.98 0.97 0.34 0.98
noise 42 2173 0.93 0.02 0.69 0.76 0.77 0.78 0.98 0.97 0.97 0.34 0.98
80s 117 2098 0.95 0.06 0.17 0.68 0.68 0.69 0.95 0.88 0.92 0.05 0.95
pop 479 1736 0.75 0.22 0.29 0.68 0.68 0.69 0.78 0.78 0.76 0.22 0.78
female 345 1870 0.8 0.16 0.29 0.8 0.8 0.79 0.85 0.83 0.82 0.15 0.85
slow 159 2056 0.93 0.07 0.32 0.65 0.69 0.68 0.92 0.93 0.9 0.07 0.93
funk 37 2178 0.98 0.02 0.43 0.71 0.71 0.7 0.98 0.96 0.97 0.34 0.98
keyboard 42 2173 0.98 0.02 0.17 0.59 0.56 0.53 0.98 0.98 0.96 0.02 0.98
electronica 166 2049 0.92 0.08 0.42 0.69 0.73 0.73 0.93 0.89 0.89 0.07 0.93
rock 664 1551 0.72 0.3 0.49 0.78 0.79 0.78 0.71 0.7 0.81 0.3 0.58
vocal 270 1945 0.88 0.13 0.19 0.59 0.61 0.6 0.88 0.87 0.81 0.12 0.88
acoustic 38 2177 0.98 0.02 0.45 0.68 0.68 0.66 0.98 0.94 0.97 0.66 0.98
hip hop 152 2063 0.93 0.07 0.69 0.83 0.87 0.87 0.94 0.93 0.94 0.07 0.89
guitar 849 1366 0.63 0.39 0.52 0.72 0.73 0.73 0.64 0.62 0.72 0.38 0.52
bass 420 1795 0.81 0.19 0.22 0.56 0.57 0.59 0.81 0.78 0.71 0.19 0.81
british 82 2133 0.96 0.04 0.24 0.69 0.7 0.67 0.96 0.96 0.93 0.04 0.96
dance 324 1891 0.86 0.15 0.31 0.75 0.8 0.79 0.85 0.85 0.87 0.15 0.81
r&b 38 2177 0.98 0.02 0.66 0.75 0.76 0.74 0.98 0.98 0.97 0.66 0.98
electronic 485 1730 0.8 0.22 0.42 0.71 0.76 0.76 0.8 0.78 0.79 0.22 0.69
drum 927 1288 0.58 0.42 0.43 0.58 0.58 0.59 0.52 0.58 0.58 0.42 0.53
soft 62 2153 0.97 0.03 0.32 0.68 0.69 0.68 0.97 0.97 0.95 0.03 0.97
punk 51 2164 0.98 0.03 0.6 0.8 0.81 0.8 0.98 0.98 0.97 0.02 0.98
country 74 2141 0.92 0.04 0.22 0.73 0.73 0.75 0.95 0.97 0.94 0.03 0.97
drum machine 89 2126 0.96 0.04 0.45 0.65 0.69 0.69 0.96 0.96 0.94 0.04 0.96
voice 148 2067 0.93 0.07 0.12 0.52 0.54 0.54 0.93 0.63 0.88 0.07 0.93
quiet 43 2172 0.98 0.02 0.83 0.84 0.85 0.84 0.98 0.95 0.97 0.02 0.98
distortion 60 2155 0.97 0.03 0.54 0.77 0.78 0.77 0.97 0.97 0.95 0.03 0.97
synth 480 1735 0.79 0.22 0.32 0.66 0.67 0.67 0.78 0.78 0.74 0.22 0.77
rap 157 2058 0.93 0.07 0.65 0.86 0.89 0.89 0.93 0.93 0.95 0.07 0.89
techno 248 1967 0.89 0.11 0.56 0.74 0.8 0.8 0.89 0.89 0.89 0.11 0.86
organ 35 2180 0.98 0.02 0.39 0.55 0.54 0.52 0.98 0.97 0.97 0.34 0.98
trumpet 39 2176 0.98 0.02 0.39 0.77 0.76 0.73 0.98 0.98 0.97 0.34 0.98
MEAN 198.76 2016.24 0.91 0.09 0.43 0.71 0.73 0.72 0.9 0.89 0.9 0.17 0.9

download these results as csv

Summary AUC-ROC Tag (Average Across All Folds)

Tag BBE1 BBE2 BBE3 ME1 ME2 ME3 LB TB TTKV
metal 0.78 0.3 0.84 0.85 0.85 0.85 0.87 0.48 0.52
ambient 0.76 0.76 0.76 0.74 0.79 0.76 0.82 0.5 0.5
fast 0.58 0.38 0.77 0.73 0.77 0.78 0.75 0.55 0.52
solo 0.57 0.6 0.69 0.74 0.71 0.7 0.64 0.59 0.6
jazz 0.78 0.69 0.85 0.89 0.9 0.89 0.84 0.49 0.52
instrumental 0.53 0.6 0.6 0.66 0.69 0.67 0.57 0.46 0.47
horns 0.57 0.48 0.56 0.63 0.64 0.57 0.64 0.38 0.38
house 0.68 0.4 0.79 0.83 0.87 0.84 0.84 0.49 0.49
male 0.6 0.44 0.75 0.79 0.8 0.8 0.69 0.49 0.49
beat 0.68 0.43 0.79 0.79 0.82 0.81 0.84 0.5 0.49
strings 0.67 0.58 0.74 0.79 0.82 0.82 0.84 0.48 0.5
saxophone 0.78 0.7 0.81 0.9 0.9 0.88 0.84 0.52 0.5
piano 0.67 0.68 0.77 0.75 0.78 0.78 0.81 0.49 0.5
loud 0.82 0.24 0.87 0.86 0.86 0.87 0.87 0.49 0.5
noise 0.73 0.4 0.83 0.75 0.77 0.77 0.8 0.45 0.46
80s 0.56 0.5 0.67 0.69 0.7 0.73 0.76 0.5 0.5
pop 0.64 0.43 0.73 0.74 0.75 0.76 0.74 0.49 0.48
female 0.59 0.48 0.75 0.86 0.85 0.84 0.75 0.51 0.5
slow 0.72 0.63 0.79 0.7 0.75 0.74 0.82 0.52 0.5
funk 0.66 0.42 0.67 0.81 0.81 0.79 0.76 0.58 0.52
keyboard 0.51 0.6 0.58 0.55 0.54 0.51 0.62 0.5 0.53
electronica 0.65 0.5 0.79 0.76 0.78 0.77 0.77 0.45 0.46
rock 0.75 0.37 0.86 0.85 0.86 0.86 0.87 0.51 0.5
vocal 0.55 0.47 0.57 0.63 0.65 0.65 0.66 0.49 0.48
acoustic 0.67 0.62 0.73 0.74 0.73 0.72 0.84 0.52 0.53
hip hop 0.82 0.37 0.89 0.92 0.93 0.92 0.93 0.54 0.52
guitar 0.69 0.42 0.81 0.79 0.8 0.8 0.79 0.5 0.5
bass 0.5 0.41 0.58 0.55 0.57 0.61 0.58 0.5 0.5
british 0.61 0.45 0.57 0.69 0.7 0.69 0.72 0.51 0.5
dance 0.74 0.41 0.87 0.8 0.85 0.84 0.86 0.49 0.49
r&b 0.7 0.45 0.74 0.8 0.81 0.81 0.83 0.48 0.45
electronic 0.71 0.48 0.81 0.78 0.81 0.8 0.77 0.49 0.49
drum 0.57 0.42 0.63 0.61 0.61 0.61 0.61 0.51 0.49
soft 0.73 0.67 0.8 0.75 0.79 0.74 0.83 0.49 0.49
punk 0.71 0.39 0.67 0.87 0.84 0.83 0.84 0.53 0.53
country 0.67 0.5 0.81 0.75 0.75 0.75 0.82 0.51 0.49
drum machine 0.64 0.47 0.76 0.73 0.76 0.75 0.78 0.5 0.5
voice 0.54 0.46 0.52 0.52 0.53 0.53 0.6 0.53 0.51
quiet 0.7 0.75 0.8 0.86 0.89 0.86 0.93 0.5 0.48
distortion 0.66 0.41 0.78 0.77 0.79 0.8 0.82 0.46 0.46
synth 0.59 0.51 0.72 0.71 0.72 0.72 0.68 0.5 0.5
rap 0.85 0.36 0.93 0.94 0.95 0.94 0.93 0.53 0.53
techno 0.75 0.41 0.86 0.82 0.87 0.86 0.85 0.46 0.47
organ 0.51 0.62 0.54 0.5 0.46 0.51 0.54 0.54 0.55
trumpet 0.64 0.61 0.7 0.8 0.79 0.73 0.71 0.53 0.51
MEAN 0.66 0.5 0.74 0.75 0.77 0.76 0.77 0.5 0.5

download these results as csv

Friedman test results

AUC-ROC Tag Friedman test

The following table and plot show the results of Friedman's ANOVA with Tukey-Kramer multiple comparisons computed over the Area Under the ROC curve (AUC-ROC) for each tag in the test, averaged over all folds.

TeamID TeamID Lowerbound Mean Upperbound Significance
BBE1 BBE2 -1.8130 -0.0222 1.7686 FALSE
BBE1 BBE3 -1.1686 0.6222 2.4130 FALSE
BBE1 ME1 -0.3686 1.4222 3.2130 FALSE
BBE1 ME2 -0.6130 1.1778 2.9686 FALSE
BBE1 ME3 1.6092 3.4000 5.1908 TRUE
BBE1 LB 3.3203 5.1111 6.9019 TRUE
BBE1 TB 3.4537 5.2444 7.0352 TRUE
BBE1 TTKV 3.4537 5.2444 7.0352 TRUE
BBE2 BBE3 -1.1463 0.6444 2.4352 FALSE
BBE2 ME1 -0.3463 1.4444 3.2352 FALSE
BBE2 ME2 -0.5908 1.2000 2.9908 FALSE
BBE2 ME3 1.6314 3.4222 5.2130 TRUE
BBE2 LB 3.3425 5.1333 6.9241 TRUE
BBE2 TB 3.4759 5.2667 7.0575 TRUE
BBE2 TTKV 3.4759 5.2667 7.0575 TRUE
BBE3 ME1 -0.9908 0.8000 2.5908 FALSE
BBE3 ME2 -1.2352 0.5556 2.3463 FALSE
BBE3 ME3 0.9870 2.7778 4.5686 TRUE
BBE3 LB 2.6981 4.4889 6.2797 TRUE
BBE3 TB 2.8314 4.6222 6.4130 TRUE
BBE3 TTKV 2.8314 4.6222 6.4130 TRUE
ME1 ME2 -2.0352 -0.2444 1.5463 FALSE
ME1 ME3 0.1870 1.9778 3.7686 TRUE
ME1 LB 1.8981 3.6889 5.4797 TRUE
ME1 TB 2.0314 3.8222 5.6130 TRUE
ME1 TTKV 2.0314 3.8222 5.6130 TRUE
ME2 ME3 0.4314 2.2222 4.0130 TRUE
ME2 LB 2.1425 3.9333 5.7241 TRUE
ME2 TB 2.2759 4.0667 5.8575 TRUE
ME2 TTKV 2.2759 4.0667 5.8575 TRUE
ME3 LB -0.0797 1.7111 3.5019 FALSE
ME3 TB 0.0537 1.8444 3.6352 TRUE
ME3 TTKV 0.0537 1.8444 3.6352 TRUE
LB TB -1.6575 0.1333 1.9241 FALSE
LB TTKV -1.6575 0.1333 1.9241 FALSE
TB TTKV -1.7908 0.0000 1.7908 FALSE

download these results as csv

2008 affinity.auc roc tag.friedman.tukeykramerhsd.png

AUC-ROC Track Friedman test

The following table and plot show the results of Friedman's ANOVA with Tukey-Kramer multiple comparisons computed over the Area Under the ROC curve (AUC-ROC) for each track in the test. Each track appears in exactly once over all three folds of the test. However, we are uncertain if these measurements are truly independent as an multiple tracks from each artist are used.

TeamID TeamID Lowerbound Mean Upperbound Significance
BBE1 BBE2 0.0026 0.2567 0.5107 TRUE
BBE1 BBE3 0.1900 0.4440 0.6981 TRUE
BBE1 ME1 0.4320 0.6860 0.9401 TRUE
BBE1 ME2 0.6019 0.8560 1.1100 TRUE
BBE1 ME3 0.7392 0.9932 1.2473 TRUE
BBE1 LB 0.8712 1.1253 1.3793 TRUE
BBE1 TB 1.6886 1.9427 2.1967 TRUE
BBE1 TTKV 4.0410 4.2950 4.5491 TRUE
BBE2 BBE3 -0.0667 0.1874 0.4414 FALSE
BBE2 ME1 0.1753 0.4293 0.6834 TRUE
BBE2 ME2 0.3453 0.5993 0.8534 TRUE
BBE2 ME3 0.4825 0.7366 0.9906 TRUE
BBE2 LB 0.6146 0.8686 1.1227 TRUE
BBE2 TB 1.4320 1.6860 1.9401 TRUE
BBE2 TTKV 3.7843 4.0384 4.2924 TRUE
BBE3 ME1 -0.0121 0.2420 0.4960 FALSE
BBE3 ME2 0.1579 0.4120 0.6660 TRUE
BBE3 ME3 0.2952 0.5492 0.8033 TRUE
BBE3 LB 0.4272 0.6813 0.9353 TRUE
BBE3 TB 1.2446 1.4986 1.7527 TRUE
BBE3 TTKV 3.5970 3.8510 4.1051 TRUE
ME1 ME2 -0.0841 0.1700 0.4240 FALSE
ME1 ME3 0.0532 0.3072 0.5613 TRUE
ME1 LB 0.1852 0.4393 0.6933 TRUE
ME1 TB 1.0026 1.2567 1.5107 TRUE
ME1 TTKV 3.3550 3.6090 3.8631 TRUE
ME2 ME3 -0.1168 0.1372 0.3913 FALSE
ME2 LB 0.0153 0.2693 0.5233 TRUE
ME2 TB 0.8326 1.0867 1.3407 TRUE
ME2 TTKV 3.1850 3.4391 3.6931 TRUE
ME3 LB -0.1220 0.1321 0.3861 FALSE
ME3 TB 0.6954 0.9494 1.2035 TRUE
ME3 TTKV 3.0478 3.3018 3.5559 TRUE
LB TB 0.5633 0.8174 1.0714 TRUE
LB TTKV 2.9157 3.1698 3.4238 TRUE
TB TTKV 2.0983 2.3524 2.6064 TRUE

download these results as csv

2008 affinity.auc roc track.friedman.tukeykramerhsd.png

Tag Classification Accuracy Friedman test

The following table and plot show the results of Friedman's ANOVA with Tukey-Kramer multiple comparisons computed over the classification accuracy for each tag in the test, averaged over all folds.

TeamID TeamID Lowerbound Mean Upperbound Significance
BBE1 BBE2 -2.5705 -0.3333 1.9038 FALSE
BBE1 BBE3 -2.2594 -0.0222 2.2149 FALSE
BBE1 ME1 -1.1149 1.1222 3.3594 FALSE
BBE1 ME2 -1.6149 0.6222 2.8594 FALSE
BBE1 ME3 0.8628 3.1000 5.3372 TRUE
BBE1 GP1 1.2406 3.4778 5.7149 TRUE
BBE1 GP2 1.9517 4.1889 6.4261 TRUE
BBE1 LB 3.7962 6.0333 8.2705 TRUE
BBE1 TB 5.4628 7.7000 9.9372 TRUE
BBE1 TTKV 4.9962 7.2333 9.4705 TRUE
BBE2 BBE3 -1.9261 0.3111 2.5483 FALSE
BBE2 ME1 -0.7816 1.4556 3.6927 FALSE
BBE2 ME2 -1.2816 0.9556 3.1927 FALSE
BBE2 ME3 1.1962 3.4333 5.6705 TRUE
BBE2 GP1 1.5739 3.8111 6.0483 TRUE
BBE2 GP2 2.2851 4.5222 6.7594 TRUE
BBE2 LB 4.1295 6.3667 8.6038 TRUE
BBE2 TB 5.7962 8.0333 10.2705 TRUE
BBE2 TTKV 5.3295 7.5667 9.8038 TRUE
BBE3 ME1 -1.0927 1.1444 3.3816 FALSE
BBE3 ME2 -1.5927 0.6444 2.8816 FALSE
BBE3 ME3 0.8851 3.1222 5.3594 TRUE
BBE3 GP1 1.2628 3.5000 5.7372 TRUE
BBE3 GP2 1.9739 4.2111 6.4483 TRUE
BBE3 LB 3.8184 6.0556 8.2927 TRUE
BBE3 TB 5.4851 7.7222 9.9594 TRUE
BBE3 TTKV 5.0184 7.2556 9.4927 TRUE
ME1 ME2 -2.7372 -0.5000 1.7372 FALSE
ME1 ME3 -0.2594 1.9778 4.2149 FALSE
ME1 GP1 0.1184 2.3556 4.5927 TRUE
ME1 GP2 0.8295 3.0667 5.3038 TRUE
ME1 LB 2.6739 4.9111 7.1483 TRUE
ME1 TB 4.3406 6.5778 8.8149 TRUE
ME1 TTKV 3.8739 6.1111 8.3483 TRUE
ME2 ME3 0.2406 2.4778 4.7149 TRUE
ME2 GP1 0.6184 2.8556 5.0927 TRUE
ME2 GP2 1.3295 3.5667 5.8038 TRUE
ME2 LB 3.1739 5.4111 7.6483 TRUE
ME2 TB 4.8406 7.0778 9.3149 TRUE
ME2 TTKV 4.3739 6.6111 8.8483 TRUE
ME3 GP1 -1.8594 0.3778 2.6149 FALSE
ME3 GP2 -1.1483 1.0889 3.3261 FALSE
ME3 LB 0.6962 2.9333 5.1705 TRUE
ME3 TB 2.3628 4.6000 6.8372 TRUE
ME3 TTKV 1.8962 4.1333 6.3705 TRUE
GP1 GP2 -1.5261 0.7111 2.9483 FALSE
GP1 LB 0.3184 2.5556 4.7927 TRUE
GP1 TB 1.9851 4.2222 6.4594 TRUE
GP1 TTKV 1.5184 3.7556 5.9927 TRUE
GP2 LB -0.3927 1.8444 4.0816 FALSE
GP2 TB 1.2739 3.5111 5.7483 TRUE
GP2 TTKV 0.8073 3.0444 5.2816 TRUE
LB TB -0.5705 1.6667 3.9038 FALSE
LB TTKV -1.0372 1.2000 3.4372 FALSE
TB TTKV -2.7038 -0.4667 1.7705 FALSE

download these results as csv

2008 binary accuracy.friedman.tukeykramerhsd.png

Tag F-measure Friedman test

The following table and plot show the results of Friedman's ANOVA with Tukey-Kramer multiple comparisons computed over the F-measure for each tag in the test, averaged over all folds.

TeamID TeamID Lowerbound Mean Upperbound Significance
BBE1 BBE2 -2.8274 -0.6000 1.6274 FALSE
BBE1 BBE3 -2.1163 0.1111 2.3385 FALSE
BBE1 ME1 -1.8274 0.4000 2.6274 FALSE
BBE1 ME2 -0.5830 1.6444 3.8719 FALSE
BBE1 ME3 0.7948 3.0222 5.2496 TRUE
BBE1 GP1 1.8392 4.0667 6.2941 TRUE
BBE1 GP2 3.5504 5.7778 8.0052 TRUE
BBE1 LB 4.5726 6.8000 9.0274 TRUE
BBE1 TB 4.2948 6.5222 8.7496 TRUE
BBE1 TTKV 4.2504 6.4778 8.7052 TRUE
BBE2 BBE3 -1.5163 0.7111 2.9385 FALSE
BBE2 ME1 -1.2274 1.0000 3.2274 FALSE
BBE2 ME2 0.0170 2.2444 4.4719 TRUE
BBE2 ME3 1.3948 3.6222 5.8496 TRUE
BBE2 GP1 2.4392 4.6667 6.8941 TRUE
BBE2 GP2 4.1504 6.3778 8.6052 TRUE
BBE2 LB 5.1726 7.4000 9.6274 TRUE
BBE2 TB 4.8948 7.1222 9.3496 TRUE
BBE2 TTKV 4.8504 7.0778 9.3052 TRUE
BBE3 ME1 -1.9385 0.2889 2.5163 FALSE
BBE3 ME2 -0.6941 1.5333 3.7608 FALSE
BBE3 ME3 0.6837 2.9111 5.1385 TRUE
BBE3 GP1 1.7281 3.9556 6.1830 TRUE
BBE3 GP2 3.4392 5.6667 7.8941 TRUE
BBE3 LB 4.4615 6.6889 8.9163 TRUE
BBE3 TB 4.1837 6.4111 8.6385 TRUE
BBE3 TTKV 4.1392 6.3667 8.5941 TRUE
ME1 ME2 -0.9830 1.2444 3.4719 FALSE
ME1 ME3 0.3948 2.6222 4.8496 TRUE
ME1 GP1 1.4392 3.6667 5.8941 TRUE
ME1 GP2 3.1504 5.3778 7.6052 TRUE
ME1 LB 4.1726 6.4000 8.6274 TRUE
ME1 TB 3.8948 6.1222 8.3496 TRUE
ME1 TTKV 3.8504 6.0778 8.3052 TRUE
ME2 ME3 -0.8496 1.3778 3.6052 FALSE
ME2 GP1 0.1948 2.4222 4.6496 TRUE
ME2 GP2 1.9059 4.1333 6.3608 TRUE
ME2 LB 2.9281 5.1556 7.3830 TRUE
ME2 TB 2.6504 4.8778 7.1052 TRUE
ME2 TTKV 2.6059 4.8333 7.0608 TRUE
ME3 GP1 -1.1830 1.0444 3.2719 FALSE
ME3 GP2 0.5281 2.7556 4.9830 TRUE
ME3 LB 1.5504 3.7778 6.0052 TRUE
ME3 TB 1.2726 3.5000 5.7274 TRUE
ME3 TTKV 1.2281 3.4556 5.6830 TRUE
GP1 GP2 -0.5163 1.7111 3.9385 FALSE
GP1 LB 0.5059 2.7333 4.9608 TRUE
GP1 TB 0.2281 2.4556 4.6830 TRUE
GP1 TTKV 0.1837 2.4111 4.6385 TRUE
GP2 LB -1.2052 1.0222 3.2496 FALSE
GP2 TB -1.4830 0.7444 2.9719 FALSE
GP2 TTKV -1.5274 0.7000 2.9274 FALSE
LB TB -2.5052 -0.2778 1.9496 FALSE
LB TTKV -2.5496 -0.3222 1.9052 FALSE
TB TTKV -2.2719 -0.0444 2.1830 FALSE

download these results as csv

2008 binary fmeasure.friedman.tukeykramerhsd.png


Beta-Binomial test results

Accuracy on positive examples Beta-Binomial results

The following table and plot show the results of simulations from the Beta-Binomial model using the accuracy of each algorithm's classification only on the positive examples. It only shows the relative proportion of true positives and false negatives, and should be considered with the classification accuracy on the negative examples. The image shows the estimate of the overall performance with 95% confidence intervals.


Tag BBE1 BBE2 BBE3 ME1 ME2 ME3 GP1 GP2 LB TB TTKV
80s 0.006 1.000 0.930 0.597 0.587 0.629 0.000 0.171 0.225 1.000 0.000
acoustic 0.000 1.000 0.737 0.650 0.617 0.639 0.000 0.000 0.185 0.346 0.000
ambient 0.000 0.988 0.420 0.552 0.620 0.639 0.000 0.083 0.273 1.000 0.000
bass 0.000 1.000 0.954 0.499 0.539 0.546 0.000 0.072 0.228 1.000 0.000
beat 0.000 1.000 0.800 0.747 0.787 0.738 0.000 0.035 0.250 1.000 0.000
british 0.008 1.000 0.842 0.613 0.628 0.658 0.000 0.000 0.107 1.000 0.000
country 0.177 1.000 0.979 0.601 0.619 0.606 0.052 0.000 0.148 1.000 0.000
dance 0.040 1.000 0.990 0.729 0.747 0.725 0.041 0.000 0.540 1.000 0.042
distortion 0.000 1.000 0.822 0.675 0.626 0.640 0.000 0.000 0.133 1.000 0.000
drum 0.008 1.000 1.000 0.609 0.588 0.573 0.358 0.000 0.495 1.000 0.281
drum machine 0.000 1.000 0.893 0.662 0.693 0.692 0.000 0.000 0.217 1.000 0.000
electronic 0.196 0.998 0.968 0.687 0.702 0.690 0.190 0.000 0.519 1.000 0.157
electronica 0.059 1.000 0.943 0.682 0.724 0.682 0.000 0.036 0.283 1.000 0.000
fast 0.000 1.000 0.889 0.697 0.695 0.740 0.000 0.019 0.223 1.000 0.000
female 0.121 1.000 0.949 0.755 0.716 0.727 0.006 0.023 0.437 1.000 0.004
funk 0.000 1.000 0.748 0.726 0.754 0.697 0.000 0.021 0.029 0.653 0.000
guitar 0.121 1.000 0.987 0.722 0.715 0.723 0.158 0.000 0.643 1.000 0.368
hip hop 0.123 1.000 0.896 0.831 0.856 0.817 0.241 0.000 0.582 1.000 0.044
horns 0.000 1.000 0.805 0.539 0.539 0.481 0.000 0.057 0.020 0.659 0.000
house 0.000 1.000 0.876 0.819 0.775 0.742 0.000 0.064 0.199 1.000 0.000
instrumental 0.000 1.000 0.727 0.587 0.621 0.591 0.312 0.036 0.076 1.000 0.000
jazz 0.052 1.000 0.915 0.702 0.742 0.723 0.043 0.000 0.407 1.000 0.017
keyboard 0.000 1.000 0.907 0.450 0.547 0.462 0.000 0.000 0.000 1.000 0.000
loud 0.000 1.000 0.802 0.696 0.666 0.696 0.000 0.000 0.165 0.688 0.000
male 0.125 1.000 0.997 0.714 0.717 0.722 0.011 0.000 0.497 1.000 0.197
metal 0.000 1.000 0.741 0.774 0.657 0.673 0.000 0.104 0.179 0.341 0.000
noise 0.240 0.968 0.698 0.581 0.501 0.544 0.000 0.000 0.161 0.659 0.000
organ 0.000 1.000 0.675 0.410 0.321 0.333 0.000 0.000 0.000 0.659 0.000
piano 0.056 1.000 0.673 0.671 0.721 0.747 0.000 0.000 0.338 1.000 0.000
pop 0.098 1.000 0.985 0.667 0.690 0.713 0.023 0.000 0.450 1.000 0.000
punk 0.000 1.000 0.709 0.752 0.707 0.632 0.000 0.000 0.305 1.000 0.000
quiet 0.165 0.865 0.544 0.624 0.751 0.703 0.000 0.128 0.254 1.000 0.000
r&b 0.025 1.000 0.560 0.600 0.647 0.696 0.025 0.000 0.179 0.341 0.000
rap 0.461 1.000 0.677 0.861 0.876 0.852 0.000 0.000 0.647 1.000 0.052
rock 0.241 1.000 0.981 0.745 0.747 0.733 0.049 0.011 0.684 1.000 0.244
saxophone 0.000 1.000 0.846 0.795 0.821 0.724 0.000 0.000 0.339 1.000 0.000
slow 0.000 1.000 0.969 0.637 0.689 0.675 0.080 0.000 0.329 1.000 0.000
soft 0.000 1.000 0.916 0.741 0.805 0.690 0.094 0.000 0.194 1.000 0.000
solo 0.000 1.000 0.979 0.612 0.594 0.594 0.000 0.000 0.097 1.000 0.000
strings 0.077 1.000 0.846 0.711 0.735 0.777 0.000 0.000 0.219 1.000 0.000
synth 0.016 1.000 0.974 0.677 0.636 0.623 0.000 0.002 0.395 1.000 0.020
techno 0.000 1.000 0.920 0.764 0.800 0.747 0.000 0.000 0.498 1.000 0.016
trumpet 0.000 1.000 0.809 0.623 0.686 0.580 0.000 0.000 0.141 0.653 0.000
vocal 0.000 1.000 0.943 0.613 0.631 0.602 0.000 0.000 0.219 1.000 0.000
voice 0.000 1.000 0.966 0.531 0.592 0.516 0.000 0.366 0.087 1.000 0.000
overall 0.000 1.000 0.890 0.669 0.681 0.664 0.000 0.000 0.226 1.000 0.000

download these results as csv


File:Binary per fold positive example Accuracy.png


The plots for each tag are more interesting and the 95% confidence intervals are much tighter. Since there are so many of them, it is difficult to post them to the wiki. You can download a tar.gz zip file containing all of them here.

Accuracy on negative examples Beta-Binomial results

The following table and plot show the results of simulations from the Beta-Binomial model using the accuracy of each algorithm's classification only on the negative examples. It only shows the relative proportion of true negatives and false positives, and should be considered with the classification accuracy on the positive examples. The image shows the estimate of the overall performance with 95% confidence intervals.


Tag BBE1 BBE2 BBE3 ME1 ME2 ME3 GP1 GP2 LB TB TTKV
80s 1.000 0.003 0.129 0.682 0.684 0.697 1.000 0.921 0.957 0.000 1.000
acoustic 0.998 0.003 0.457 0.681 0.684 0.664 1.000 0.958 0.986 0.653 1.000
ambient 1.000 0.003 0.873 0.746 0.806 0.748 0.999 0.968 0.980 0.000 1.000
bass 1.000 0.003 0.055 0.572 0.579 0.602 1.000 0.942 0.821 0.000 1.000
beat 1.000 0.003 0.549 0.701 0.742 0.748 1.000 0.955 0.951 0.000 1.000
british 0.998 0.003 0.214 0.694 0.700 0.676 1.000 1.000 0.965 0.000 1.000
country 0.951 0.003 0.197 0.732 0.733 0.758 0.981 1.000 0.969 0.000 1.000
dance 0.996 0.003 0.192 0.748 0.802 0.794 0.991 1.000 0.923 0.000 0.944
distortion 1.000 0.003 0.525 0.775 0.784 0.777 1.000 1.000 0.976 0.000 1.000
drum 0.996 0.003 0.023 0.564 0.572 0.595 0.648 1.000 0.640 0.000 0.705
drum machine 1.000 0.003 0.428 0.653 0.690 0.684 1.000 1.000 0.968 0.000 1.000
electronic 0.966 0.003 0.279 0.714 0.768 0.772 0.954 1.000 0.867 0.000 0.833
electronica 0.991 0.003 0.390 0.684 0.725 0.726 1.000 0.960 0.941 0.000 1.000
fast 1.000 0.003 0.412 0.680 0.693 0.696 0.999 0.967 0.960 0.000 1.000
female 0.930 0.003 0.177 0.807 0.817 0.804 0.999 0.985 0.896 0.000 0.999
funk 1.000 0.003 0.426 0.708 0.712 0.696 1.000 0.982 0.984 0.347 1.000
guitar 0.939 0.003 0.222 0.711 0.730 0.730 0.950 1.000 0.777 0.000 0.617
hip hop 0.994 0.003 0.675 0.831 0.866 0.869 0.988 1.000 0.968 0.000 0.955
horns 1.000 0.003 0.303 0.660 0.674 0.644 1.000 0.935 0.982 0.342 1.000
house 1.000 0.003 0.413 0.756 0.797 0.756 1.000 0.955 0.975 0.000 1.000
instrumental 1.000 0.003 0.374 0.631 0.648 0.638 0.688 0.951 0.955 0.000 1.000
jazz 0.991 0.003 0.465 0.846 0.865 0.867 1.000 1.000 0.960 0.000 0.981
keyboard 1.000 0.003 0.163 0.590 0.557 0.530 1.000 1.000 0.981 0.000 1.000
loud 1.000 0.003 0.552 0.813 0.825 0.814 1.000 1.000 0.986 0.312 1.000
male 0.913 0.003 0.048 0.721 0.740 0.744 0.989 1.000 0.746 0.000 0.796
metal 1.000 0.003 0.647 0.788 0.791 0.785 1.000 0.969 0.987 0.659 1.000
noise 0.949 0.003 0.697 0.760 0.778 0.784 0.998 0.989 0.984 0.342 1.000
organ 1.000 0.003 0.393 0.555 0.547 0.527 1.000 0.991 0.984 0.341 1.000
piano 0.960 0.003 0.421 0.655 0.679 0.692 1.000 1.000 0.940 0.000 1.000
pop 0.931 0.003 0.102 0.680 0.683 0.682 0.988 1.000 0.849 0.000 1.000
punk 1.000 0.003 0.602 0.802 0.816 0.807 1.000 1.000 0.983 0.000 1.000
quiet 0.994 0.003 0.842 0.839 0.853 0.844 1.000 0.967 0.985 0.000 0.999
r&b 0.994 0.003 0.652 0.751 0.756 0.734 0.999 1.000 0.986 0.658 1.000
rap 0.967 0.003 0.639 0.853 0.885 0.887 1.000 1.000 0.972 0.000 0.950
rock 0.922 0.003 0.286 0.788 0.806 0.800 0.986 0.991 0.866 0.000 0.721
saxophone 1.000 0.003 0.455 0.852 0.852 0.841 1.000 0.985 0.978 0.000 0.999
slow 1.000 0.003 0.265 0.652 0.687 0.687 0.981 1.000 0.948 0.000 1.000
soft 1.000 0.003 0.302 0.677 0.690 0.684 0.990 1.000 0.977 0.000 1.000
solo 1.000 0.003 0.165 0.740 0.725 0.737 1.000 1.000 0.978 0.000 1.000
strings 0.969 0.003 0.466 0.759 0.754 0.771 0.998 1.000 0.980 0.000 1.000
synth 0.997 0.003 0.135 0.655 0.677 0.679 1.000 0.996 0.834 0.000 0.981
techno 1.000 0.003 0.526 0.737 0.799 0.798 1.000 1.000 0.937 0.000 0.973
trumpet 1.000 0.003 0.386 0.775 0.760 0.735 1.000 0.996 0.984 0.347 1.000
vocal 1.000 0.003 0.081 0.584 0.608 0.603 1.000 0.995 0.892 0.000 1.000
voice 1.000 0.003 0.054 0.527 0.540 0.541 1.000 0.654 0.935 0.000 1.000
overall 0.999 0.003 0.346 0.719 0.737 0.731 1.000 1.000 0.950 0.000 1.000

download these results as csv


File:Binary per fold negative example Accuracy.png


The plots for each tag are more interesting and the 95% confidence intervals are much tighter. Since there are so many of them, it is difficult to post them to the wiki. You can download a tar.gz file containing all of them here.

Assorted Results Files for Download

AUC-ROC Clip Data

(Too large for easy Wiki viewing)
tag.affinity_clip_AUC_ROC.csv

CSV Files Without Rounding (Averaged across folds)

tag.affinity.clip.auc.roc.csv
tag.affinity.clip.auc.roc.csv
tag.binary.accuracy.csv
tag.binary.fmeasure.csv
tag.binary.negative.example.accuracy.csv
tag.binary.positive.example.accuracy.csv

CSV Files Without Rounding (Fold information)

tag.binary.per.fold.positive.example.accuracy.csv
tag.binary.per.fold.negative.example.accuracy.csv
tag.binary.per.fold.fmeasure.csv
tag.binary.per.fold.accuracy.csv
tag.affinity.tag.per.fold.auc.roc.csv

Results By Algorithm

(.tar.gz)
LB = L. Barrington, D. Turnbull, G. Lanckriet
BBE 1 = T. Bertin-Mahieux, Y. Bengio, D. Eck (KNN)
BBE 2 = T. Bertin-Mahieux, Y. Bengio, D. Eck (NNet)
BBE 3 = T. Bertin-Mahieux, D. Eck, P. Lamere, Y. Bengio (Thierry/Lamere Boosting)
TB = Bertin-Mahieux (dumb/smurf)
ME1 = M. I. Mandel, D. P. W. Ellis 1
ME2 = M. I. Mandel, D. P. W. Ellis 2
ME3 = M. I. Mandel, D. P. W. Ellis 3
GP1 = G. Peeters 1
GP2 = G. Peeters 2
TTKV = K. Trohidis, G. Tsoumakas, G. Kalliris, I. Vlahavas
.