IMPROVED MARKOV SONG-REQUESTING RECOMMENDATION ALGORITHM BASED ON USER'S INSTANT BEHAVIOUR

摘要

This paper proposes an improved Markov song-requesting recommendation algorithm based on user's instant behavior. The algorithm is based on instant behavior of user's singing to make online recommendations, and combines a first-order Markov model with collaborative filtering recommendation. Recommendation results are generated by building the transfer probability matrix between songs, and at the same time, takes into account the impact of time decay on the recommendation result. The advantage of this method is that it does not require any acoustic or semantic information of the song, which greatly reduces the amount of computation. Thus the problem of new users in collaborative filtering algorithm can be solved. This method makes full analysis of the actual behavior of choosing songs, and is closer to the actual needs of users, thus producing a higher recommendation hit ratio.