Improve the application of reinforcement learning and multi‐modal information in music sentiment analysis

Abstract

AbstractIn order to improve the effect of music sentiment analysis, this paper proposes a music sentiment classification method based on lyrics and comments. This method combines lyrics and comment texts to mine richer sentiment information, and comprehensively considers the influence of the word frequency, sentiment strength and part of speech of sentiment words on sentiment classification when constructing sentiment vectors. Moreover, it matches the lyrics of the music and the substantive words in the comment with the emotional dictionary to obtain the emotional category and emotional weight of each substantive word, and calculates the statistical value of each emotional category. In addition, this paper combines reinforcement learning and multi‐modal information technology to construct a music emotion research model.