Personalized log based recommender system for song playlist generation

Abstract

With the rapid growth of smartphones and music gadgets, various companies exploit the user needs towards music by various cloud-based music platforms. In this article log- based recommendation system is proposed, in which recommendation is done using locally stored log files. The system is capable of recommending the next most suitable song. In this research considering MP3 ID3 format for dynamic playlist generation. The results of experiments conducted on MP3 ID3 song dataset accurately generate song playlist and satisfactorily update log file for the specific user and it leads to higher performance of the system. Moreover, the demonstrated log based recommender system can be applied to various other similar domains like movies, books, news recommendation.