Music recommendation system with user-based and item-based collaborative filtering technique

Abstract

Internet and E-commerce are the generators of abundant of data, causing information Overloading. The problem of information overloading is addressed by Recommendation Systems (RS). RS can provide suggestions about a new product, movie or music etc. This paper is about Music Recommendation System, which will recommend songs to users based on their past history i.e. taste. In this paper we proposed a collaborative filtering technique based on users and items. First user-item rating matrix is used to form user clusters and item clusters. Next these clusters are used to find the most similar user cluster or most similar item cluster to a target user. Finally songs are recommended from the most similar user and item clusters. The proposed algorithm is implemented on the benchmark dataset Last.fm. Results show that the performance of proposed method is better than the most popular baseline method.