removed mentions of missing SVM models from documentation and the accuracy report

dbogdanov · dbogdanov · commit 75dfc84d9cc1 · 2014-02-13T01:27:21.000+01:00
diff --git a/doc/sphinxdoc/algorithms_overview.rst b/doc/sphinxdoc/algorithms_overview.rst
@@ -175,7 +175,6 @@ Essentia also contains the following mid- and high-level descriptors:
   * moods: happy, sad, aggressive, relaxed, acoustic, electronic, party
   * western / non-western music
   * tonal / atonal
-  * perceptual speed (ie: slow, medium, fast)
   * danceability
   * voice / instrumental
   * gender (male / female singer)
diff --git a/src/examples/svm_models/accuracies_2.0.1.html b/src/examples/svm_models/accuracies_2.0.1.html
@@ -1,3 +1,5 @@
+<h1>Classifier model accuracies (5-fold cross-validation) and confusion matrices</h1>
+
 <h1>gender</h1>
 Accuracy: 86.451022%
 <html><head>
@@ -58,12 +60,8 @@ <h1>gender</h1>
 <html><head>
                 <style type="text/css">body { font-family: sans-serif; font-size: small; } table { border: 1px solid black; border-collapse: collapse; font-size: small; } th { background-color: rgb(240, 240, 240); border: 1px solid black; text-align: left; padding: 5px; } td { border: 1px solid black; text-align: right; padding: 5px; } td.good, td.hi { background-color: #00ff33; } td.mediocre { background-color: #FF9900; color:#FFFFFF; } td.bad { background-color: #FF0000; color:#FFFFFF; } td.animate span { display:none } td.animate:hover span {display: block; position: absolute; padding: 5px; margin: 20px; z-index: 100; border: 1px solid black; color:black; background-color: AntiqueWhite; text-align: left;} .unroll { text-decoration:underline; color:blue; } .unroll:hover { cursor:pointer; } h2, h3 { padding: none; margin: 0.2em 0em 0em 0em; } #footer { text-align: right; clear: both; margin-top: 2em; border-top: dashed gray 1px; }</style>
                 </head>
-                <body><table><tr><th><h3>Predicted (%)</h3></th><td></td></tr><tr><td><table><tr><td></td><th>Cluster1</th><th>Cluster2</th><th>Cluster3</th><th>Cluster4</th><th>Cluster5</th><td></td><th>Proportion</th></tr><tr><th>Cluster1</th><td class="hi animate">58.62 <span>34 Cluster1 (out of 58) classified as Cluster1</span></td><td class="bad animate">18.97 <span>11 Cluster1 (out of 58) classified as Cluster2</span></td><td class="bad animate">10.34 <span>6 Cluster1 (out of 58) classified as Cluster3</span></td><td class="animate">3.45 <span>2 Cluster1 (out of 58) classified as Cluster4</span></td><td class="animate">8.62 <span>5 Cluster1 (out of 58) classified as Cluster5</span></td><th>Cluster1</th><td>21.56 %</td></tr><tr><th>Cluster2</th><td class="bad animate">29.63 <span>16 Cluster2 (out of 54) classified as Cluster1</span></td><td class="hi animate">48.15 <span>26 Cluster2 (out of 54) classified as Cluster2</span></td><td class="bad animate">16.67 <span>9 Cluster2 (out of 54) classified as Cluster3</span></td><td class="animate">5.56 <span>3 Cluster2 (out of 54) classified as Cluster4</span></td><td class="animate">0.00 <span>0 Cluster2 (out of 54) classified as Cluster5</span></td><th>Cluster2</th><td>20.07 %</td></tr><tr><th>Cluster3</th><td class="bad animate">10.81 <span>8 Cluster3 (out of 74) classified as Cluster1</span></td><td class="bad animate">13.51 <span>10 Cluster3 (out of 74) classified as Cluster2</span></td><td class="hi animate">71.62 <span>53 Cluster3 (out of 74) classified as Cluster3</span></td><td class="animate">4.05 <span>3 Cluster3 (out of 74) classified as Cluster4</span></td><td class="animate">0.00 <span>0 Cluster3 (out of 74) classified as Cluster5</span></td><th>Cluster3</th><td>27.51 %</td></tr><tr><th>Cluster4</th><td class="bad animate">15.62 <span>5 Cluster4 (out of 32) classified as Cluster1</span></td><td class="bad animate">43.75 <span>14 Cluster4 (out of 32) classified as Cluster2</span></td><td class="bad animate">12.50 <span>4 Cluster4 (out of 32) classified as Cluster3</span></td><td class="hi animate">25.00 <span>8 Cluster4 (out of 32) classified as Cluster4</span></td><td class="animate">3.12 <span>1 Cluster4 (out of 32) classified as Cluster5</span></td><th>Cluster4</th><td>11.90 %</td></tr><tr><th>Cluster5</th><td class="bad animate">25.49 <span>13 Cluster5 (out of 51) classified as Cluster1</span></td><td class="animate">1.96 <span>1 Cluster5 (out of 51) classified as Cluster2</span></td><td class="animate">0.00 <span>0 Cluster5 (out of 51) classified as Cluster3</span></td><td class="animate">1.96 <span>1 Cluster5 (out of 51) classified as Cluster4</span></td><td class="hi animate">70.59 <span>36 Cluster5 (out of 51) classified as Cluster5</span></td><th>Cluster5</th><td>18.96 %</td></tr></table></td><th><h3>Actual (%)</h3></th></tr></table><br/></body></html><h1>perceptual_speed</h1>
-Accuracy: 75.966667%
-<html><head>
-                <style type="text/css">body { font-family: sans-serif; font-size: small; } table { border: 1px solid black; border-collapse: collapse; font-size: small; } th { background-color: rgb(240, 240, 240); border: 1px solid black; text-align: left; padding: 5px; } td { border: 1px solid black; text-align: right; padding: 5px; } td.good, td.hi { background-color: #00ff33; } td.mediocre { background-color: #FF9900; color:#FFFFFF; } td.bad { background-color: #FF0000; color:#FFFFFF; } td.animate span { display:none } td.animate:hover span {display: block; position: absolute; padding: 5px; margin: 20px; z-index: 100; border: 1px solid black; color:black; background-color: AntiqueWhite; text-align: left;} .unroll { text-decoration:underline; color:blue; } .unroll:hover { cursor:pointer; } h2, h3 { padding: none; margin: 0.2em 0em 0em 0em; } #footer { text-align: right; clear: both; margin-top: 2em; border-top: dashed gray 1px; }</style>
-                </head>
-                <body><table><tr><th><h3>Predicted (%)</h3></th><td></td></tr><tr><td><table><tr><td></td><th>fast</th><th>medium</th><th>slow</th><td></td><th>Proportion</th></tr><tr><th>fast</th><td class="hi animate">80.80 <span>808 fast (out of 1000) classified as fast</span></td><td class="bad animate">13.00 <span>130 fast (out of 1000) classified as medium</span></td><td class="animate">6.20 <span>62 fast (out of 1000) classified as slow</span></td><th>fast</th><td>33.33 %</td></tr><tr><th>medium</th><td class="bad animate">13.80 <span>138 medium (out of 1000) classified as fast</span></td><td class="hi animate">70.70 <span>707 medium (out of 1000) classified as medium</span></td><td class="bad animate">15.50 <span>155 medium (out of 1000) classified as slow</span></td><th>medium</th><td>33.33 %</td></tr><tr><th>slow</th><td class="animate">7.70 <span>77 slow (out of 1000) classified as fast</span></td><td class="bad animate">15.90 <span>159 slow (out of 1000) classified as medium</span></td><td class="hi animate">76.40 <span>764 slow (out of 1000) classified as slow</span></td><th>slow</th><td>33.33 %</td></tr></table></td><th><h3>Actual (%)</h3></th></tr></table><br/></body></html><h1>genre_dortmund</h1>
+                <body><table><tr><th><h3>Predicted (%)</h3></th><td></td></tr><tr><td><table><tr><td></td><th>Cluster1</th><th>Cluster2</th><th>Cluster3</th><th>Cluster4</th><th>Cluster5</th><td></td><th>Proportion</th></tr><tr><th>Cluster1</th><td class="hi animate">58.62 <span>34 Cluster1 (out of 58) classified as Cluster1</span></td><td class="bad animate">18.97 <span>11 Cluster1 (out of 58) classified as Cluster2</span></td><td class="bad animate">10.34 <span>6 Cluster1 (out of 58) classified as Cluster3</span></td><td class="animate">3.45 <span>2 Cluster1 (out of 58) classified as Cluster4</span></td><td class="animate">8.62 <span>5 Cluster1 (out of 58) classified as Cluster5</span></td><th>Cluster1</th><td>21.56 %</td></tr><tr><th>Cluster2</th><td class="bad animate">29.63 <span>16 Cluster2 (out of 54) classified as Cluster1</span></td><td class="hi animate">48.15 <span>26 Cluster2 (out of 54) classified as Cluster2</span></td><td class="bad animate">16.67 <span>9 Cluster2 (out of 54) classified as Cluster3</span></td><td class="animate">5.56 <span>3 Cluster2 (out of 54) classified as Cluster4</span></td><td class="animate">0.00 <span>0 Cluster2 (out of 54) classified as Cluster5</span></td><th>Cluster2</th><td>20.07 %</td></tr><tr><th>Cluster3</th><td class="bad animate">10.81 <span>8 Cluster3 (out of 74) classified as Cluster1</span></td><td class="bad animate">13.51 <span>10 Cluster3 (out of 74) classified as Cluster2</span></td><td class="hi animate">71.62 <span>53 Cluster3 (out of 74) classified as Cluster3</span></td><td class="animate">4.05 <span>3 Cluster3 (out of 74) classified as Cluster4</span></td><td class="animate">0.00 <span>0 Cluster3 (out of 74) classified as Cluster5</span></td><th>Cluster3</th><td>27.51 %</td></tr><tr><th>Cluster4</th><td class="bad animate">15.62 <span>5 Cluster4 (out of 32) classified as Cluster1</span></td><td class="bad animate">43.75 <span>14 Cluster4 (out of 32) classified as Cluster2</span></td><td class="bad animate">12.50 <span>4 Cluster4 (out of 32) classified as Cluster3</span></td><td class="hi animate">25.00 <span>8 Cluster4 (out of 32) classified as Cluster4</span></td><td class="animate">3.12 <span>1 Cluster4 (out of 32) classified as Cluster5</span></td><th>Cluster4</th><td>11.90 %</td></tr><tr><th>Cluster5</th><td class="bad animate">25.49 <span>13 Cluster5 (out of 51) classified as Cluster1</span></td><td class="animate">1.96 <span>1 Cluster5 (out of 51) classified as Cluster2</span></td><td class="animate">0.00 <span>0 Cluster5 (out of 51) classified as Cluster3</span></td><td class="animate">1.96 <span>1 Cluster5 (out of 51) classified as Cluster4</span></td><td class="hi animate">70.59 <span>36 Cluster5 (out of 51) classified as Cluster5</span></td><th>Cluster5</th><td>18.96 %</td></tr></table></td><th><h3>Actual (%)</h3></th></tr></table><br/></body></html>
+                <h1>genre_dortmund</h1>
 Accuracy: 59.141494%
 <html><head>
                 <style type="text/css">body { font-family: sans-serif; font-size: small; } table { border: 1px solid black; border-collapse: collapse; font-size: small; } th { background-color: rgb(240, 240, 240); border: 1px solid black; text-align: left; padding: 5px; } td { border: 1px solid black; text-align: right; padding: 5px; } td.good, td.hi { background-color: #00ff33; } td.mediocre { background-color: #FF9900; color:#FFFFFF; } td.bad { background-color: #FF0000; color:#FFFFFF; } td.animate span { display:none } td.animate:hover span {display: block; position: absolute; padding: 5px; margin: 20px; z-index: 100; border: 1px solid black; color:black; background-color: AntiqueWhite; text-align: left;} .unroll { text-decoration:underline; color:blue; } .unroll:hover { cursor:pointer; } h2, h3 { padding: none; margin: 0.2em 0em 0em 0em; } #footer { text-align: right; clear: both; margin-top: 2em; border-top: dashed gray 1px; }</style>
@@ -98,4 +96,4 @@ <h1>gender</h1>
 <html><head>
                 <style type="text/css">body { font-family: sans-serif; font-size: small; } table { border: 1px solid black; border-collapse: collapse; font-size: small; } th { background-color: rgb(240, 240, 240); border: 1px solid black; text-align: left; padding: 5px; } td { border: 1px solid black; text-align: right; padding: 5px; } td.good, td.hi { background-color: #00ff33; } td.mediocre { background-color: #FF9900; color:#FFFFFF; } td.bad { background-color: #FF0000; color:#FFFFFF; } td.animate span { display:none } td.animate:hover span {display: block; position: absolute; padding: 5px; margin: 20px; z-index: 100; border: 1px solid black; color:black; background-color: AntiqueWhite; text-align: left;} .unroll { text-decoration:underline; color:blue; } .unroll:hover { cursor:pointer; } h2, h3 { padding: none; margin: 0.2em 0em 0em 0em; } #footer { text-align: right; clear: both; margin-top: 2em; border-top: dashed gray 1px; }</style>
                 </head>
-                <body><table><tr><th><h3>Predicted (%)</h3></th><td></td></tr><tr><td><table><tr><td></td><th>danceable</th><th>not_danceable</th><td></td><th>Proportion</th></tr><tr><th>danceable</th><td class="hi animate">93.06 <span>134 danceable (out of 144) classified as danceable</span></td><td class="animate">6.94 <span>10 danceable (out of 144) classified as not_danceable</span></td><th>danceable</th><td>64.00 %</td></tr><tr><th>not_danceable</th><td class="animate">8.64 <span>7 not_danceable (out of 81) classified as danceable</span></td><td class="hi animate">91.36 <span>74 not_danceable (out of 81) classified as not_danceable</span></td><th>not_danceable</th><td>36.00 %</td></tr></table></td><th><h3>Actual (%)</h3></th></tr></table><br/></body></html>
+                <body><table><tr><th><h3>Predicted (%)</h3></th><td></td></tr><tr><td><table><tr><td></td><th>danceable</th><th>not_danceable</th><td></td><th>Proportion</th></tr><tr><th>danceable</th><td class="hi animate">93.06 <span>134 danceable (out of 144) classified as danceable</span></td><td class="animate">6.94 <span>10 danceable (out of 144) classified as not_danceable</span></td><th>danceable</th><td>64.00 %</td></tr><tr><th>not_danceable</th><td class="animate">8.64 <span>7 not_danceable (out of 81) classified as danceable</span></td><td class="hi animate">91.36 <span>74 not_danceable (out of 81) classified as not_danceable</span></td><th>not_danceable</th><td>36.00 %</td></tr></table></td><th><h3>Actual (%)</h3></th></tr></table><br/></body></html>
diff --git a/src/examples/svm_models/model_accuracies.txt b/src/examples/svm_models/model_accuracies.txt
@@ -14,7 +14,6 @@ mood_sad            86.956522%  results/mood_sad/mood_sad40553487_0.result
 culture             93.495190%  results/culture/culture862388605_0.result
 mood_party          88.381743%  results/mood_party/mood_party810960596_0.result
 mood                58.364312%  results/mood/mood-525915621_0.result
-perceptual_speed    75.966667%  results/perceptual_speed/perceptual_speed1190097618_0.result
 genre_dortmund      59.141494%  results/genre_dortmund/genre_dortmund-1853309960_0.result
 tonal_atonal        97.674419%  results/tonal_atonal/tonal_atonal552770384_0.result
 mood_aggressive     97.500000%  results/mood_aggressive/mood_aggressive-1675548205_0.result
