Abstract
In this study, we try to predict music mood by using the online digital music database - Allmusic.com. We extracted English song lyrics from Allmusic.com. This study uses Support Vector Machine (SVM) to analyze the dataset. We calculate every token’s Log Likelihood Ratio after tokenization, removing stop-words and porter stemming. Further, we select out feature tokens as our classifying rule. Finally, we put them into LibSVM to build model, which tries to predict music mood. Findings show that the accuracy depends on lyrics is in an acceptable level. We find that mood Happy, Fiery and Drama are easy to detect. It perhaps that their emotional terms are clear. On the other hand, mood Wry, Literate, Ironic and Silly are hard to detect.
Recommended Citation
Yang, Keng-Chieh; Huang, Chia-Hwa; Yang, Conna; and Lin, Yi-Sin, "A Study on Music Mood Detection in Online Digital Music Database" (2017). PACIS 2017 Proceedings. 144.
https://aisel.aisnet.org/pacis2017/144