A hybrid music recommendation method based on music genes and collaborative filtering

With the rapid popularization and development of the Internet and Multimedia, music has become one of the most popular ways of entertainment for the public. However, in the face of excessive music information on the Internet, users often encounter the problem of music selection. Therefore, how to accurately find the music that users like from the massive music data is a difficult problem. In view of this problem, music recommendation system came into being. This paper summarizes the research results related to music recommendation, and analyzes the problems that need to be solved. Aiming at the problems existing in current music recommendation methods, this paper introduces the concept of music gene and proposes a new hybrid recommendation method.In this hybrid recommendation, the recommendation results of collaborative filtering and the recommendation results generated based on music genes are weighted and mixed, and the two methods are given different weights according to the actual situation, so as to obtain a new comprehensive recommendation result.By using this hybrid method, the impact of cold start and data sparsity problems can be effectively reduced, and the accuracy and scalability of recommendations can be improved.