Measuring the Effect of Social Network Data on Music Recommendation

Abstract

With the rapid growth of online music market, online music providers are devoting on how to recommend suitable music for their customer to fit their interests. Music recommender system is therefore has been developed and researchers are focusing on how to improve the performance of music recommender system. Nowadays, social recommender system has been discussed widely, due to the growth of social networking website. Large amount of social data can be collected and which considered can be used to improve the recommendation quality due to the characteristics of social data. Thus, it is interesting to know the performance when adopting different kind of social data into music recommender system, including "Likes", "Check-in" and "Friends" in a fans page. A series of experiments have been conducted in the paper and to measure the rating of music recommendation by considering those kinds of social data. The experiment results show that the rating is the best when the recommendation that generated by considering all three kinds of social data.