An improved recommender system based on clustering using representatives

Abstract

Recommender systems have proven to be valuable means for online users to cope up with the information overload and have become one of the most powerful and popular tools in electronic commerce. Collaborative Filtering (CF) is one of the most successful recommendation techniques that recommends by using the opinions of a community of users. However, the similarity computations associated with CF algorithms are very expensive and grow polynomially with the number of users and items in a database. To address this scalability problem, we propose a clustering based recommendation approach. Our proposed work partitions the users of the CF system using a CURE (Clustering using representatives) based data clustering algorithm and use the clusters to select the similar users of a target user. In this work, we further try to find the optimal number of clusters by using a binary search based technique. The cluster-based approach reduces the runtime of the system as we avoid similarity computations over the entire database. Experiments performed on MovieLens-1M dataset indicate that our method is efficient in reducing the runtime as well as maintains an acceptable recommendation quality.