A Hybrid Music Recommendation Model Based on Personalized Measurement and Game Theory

Music recommendation algorithms, from the perspective of real-time, can be classified into two categories: offline recommendation algorithms and online recommendation algorithms. To improve music recommendation accuracy, especially for the new music (users have no historic listening records on it), and real-time recommendation ability, and solve the interest drift problem simultaneously, we propose a hybrid music recommendation model based on personalized measurement and game theory. This model can be separated into two parts: an offline recommendation part (OFFLRP) and an online recommendation part (ONLRP). In the offline part, we emphasize users personalization. We introduce two metrics named user pursue-novelty degree (UPND) and music popularity (MP) to improve the traditional items-based collaborative filtering algorithm. In the online part, we try to solve the interest drift problem, which is a thorny problem in the offline part. We propose a novel online recommendation algorithm based on game theory. Experiments verify that the hybrid music recommendation model has higher new music recommendation accuracy, decent dynamical personalized recommendation ability, and real-time recommendation capability, and can substantially mitigate the problem of interest drift.