Abstract:

The object of this study is to improve the quality of the description of music resources, improve the efficiency of personalized music recommendations, and provide theoretical guidances for the optimization of the music recommendation system. Based on user's cognitive framework of music works, we built metadata system, and describe music works based on user's consensus. On this basis, we built user interest model that integrates group cognition and personal preferences. Taking the music of the Douban as an example, the personalized recommendation experiment was carried out to verify the effect of the model. Experimental results showed that the fusion of individual interest and group cognitive music personalized recommendation can significantly improve music recommendation accuracy.

Key words: Music recommendation, Resource annotation, Group cognition, User cognition