Review on sentiment analysis on music
Stuti Shukla, Pooja R. Khanna, Krishna Kant Agrawal · 2017
In this study, we present a review on how lyrics can be prove its usefulness in mood classification utilizing the features like linguistic lyric feature set and text stylistic feature set. From this study, one may be able to understand the concept and process of sentiment analysis on the basis of lyrics. Some psychological tools like Big five inventory and CAC scale need to be studied to understand which personality type may generate what perception to a particular song. Big five inventory gives the measure of personality type and CAC scale gives the measure of perception. Performance of lyric feature sets was measured individually as well as by combining them. Fusion methods were used to measure performance of combined feature sets. Basically, two fusion methods used were Feature concatenation and Late fusion. Additionally, examination of learning curves was done which proves that the less number of training samples were required for lyrics+audio system. The discoveries have shown the best class in lyrics sentiment analysis and musical mood classification.