Enriching Music Information Retrieval Using Emotion Detection

Music is inherently expressive of affective meaning and nondescriptive queries of Music Information Retrieval (MIR) system can show the emotion of users. In this paper, we consider the implicit emotional association between users and music to enrich MIR. We try to understand the latent emotional intent of queries via machine learning based emotion classification and compare the performance of emotion detection approaches on different feature sets. The experimental results conducted on 25 distinctive queries and 2257 songs show that our approach can be a significant exploration in processing non-descriptive music retrieval queries with interactive tags and rich lyrics.