Using emotional context from article for contextual music recommendation

• DOI: 10.1145/2502081.2502170
• Publication Date: 2013-10-22T13:42:56Z
• Authors: Chih-Ming Chen; Ming-Feng Tsai; Jen-Yu Liu; Yi-Hsuan Yang
• Keywords: 0202 electrical engineering, electronic engineering, information engineering; 02 engineering and technology

This paper proposes a context-aware approach that recommends music to user based on the user's emotional state predicted from article writes. We analyze association between user-generated text and by using real-world dataset with user, text, tripartite information collected social blogging website LiveJournal. The audio represents various perceptual dimensions of listening, including danceability, loudness, mode, tempo; consists bag-of-words three dimensional affective states within an article: valence, arousal dominance. To combine these factors for recommendation, factorization machine-based is taken. Our evaluation shows context mined articles does improve quality comparing either collaborative filtering or content-based approach.