21.01.2014 Views

improving music mood classification using lyrics, audio and social tags

improving music mood classification using lyrics, audio and social tags

improving music mood classification using lyrics, audio and social tags

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

Table 7.4 Accuracies of lyric <strong>and</strong> <strong>audio</strong> feature types for individual categories<br />

Category Content GI GI-lex ANEW Affect-lex TextStyle Audio<br />

calm 0.5905 0.5851 0.5804 0.5563 0.5708 0.5039 0.6574<br />

sad 0.6655 0.6218 0.6010 0.5441 0.5836 0.5153 0.6749<br />

glad 0.5627 0.5547 0.5600 0.5635 0.5508 0.5380 0.5882<br />

romantic 0.6866 0.6228 0.6721 0.6027 0.6333 0.5153 0.6188<br />

gleeful 0.5864 0.5763 0.5405 0.5103 0.5443 0.5670 0.6253<br />

gloomy 0.6157 0.5710 0.6124 0.5520 0.5859 0.5468 0.6178<br />

angry 0.7047 0.6362 0.6497 0.6363 0.6849 0.4924 0.5905<br />

mournful 0.6670 0.6344 0.5871 0.6058 0.6615 0.5001 0.6278<br />

dreamy 0.6143 0.5686 0.6264 0.5183 0.6269 0.5645 0.6681<br />

cheerful 0.6226 0.5633 0.5707 0.5955 0.5171 0.5105 0.5133<br />

brooding 0.5261 0.5295 0.5739 0.4985 0.5383 0.5045 0.6019<br />

aggressive 0.7966 0.7178 0.7549 0.6432 0.6746 0.5345 0.6417<br />

anxious 0.6125 0.5375 0.5750 0.5687 0.5875 0.4875 0.4875<br />

confident 0.3917 0.4429 0.4774 0.4190 0.5548 0.5083 0.5417<br />

hopeful 0.5700 0.4975 0.6025 0.5125 0.6350 0.5375 0.4000<br />

earnest 0.6125 0.6500 0.5500 0.6250 0.6000 0.6375 0.5750<br />

cynical 0.7000 0.6792 0.6375 0.4625 0.6667 0.5250 0.6292<br />

exciting 0.5833 0.5500 0.5833 0.4000 0.4667 0.5333 0.3667<br />

AVERAGE 0.6172 0.5855 0.5975 0.5452 0.5935 0.5290 0.5792<br />

The accuracies marked in bold in Table 7.4 demonstrate that <strong>lyrics</strong> <strong>and</strong> <strong>audio</strong> indeed have<br />

their respective advantages in different <strong>mood</strong> categories. Audio features significantly<br />

outperformed all lyric feature types in only one <strong>mood</strong> category: “calm.” However, lyric features<br />

achieved significantly better performances than <strong>audio</strong> in seven divergent categories: “romantic,”<br />

“angry,” “cheerful,” “aggressive,” “anxious,” “hopeful,” <strong>and</strong> “exciting.”<br />

The rest of the section presents <strong>and</strong> analyzes the most influential features of those lyric<br />

feature types that outperformed <strong>audio</strong> features in the seven aforementioned <strong>mood</strong> categories.<br />

Since the <strong>classification</strong> model used in this research was SVM with a linear kernel, the features<br />

were ranked by the same SVM models trained in the <strong>classification</strong> experiments.<br />

98

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!