A collaborative recommender system: Lexicographic consensus and web usage mining approach

Abstract

Collaborative filtering (CF) is one of the most successful and widely used methods of automated product recommendation domain. For today's large e-commerce sites, it is more recommended to use implicit ratings for constructing the customer profile. It relaxes the burden of preference specification imposed on customers. In this paper, we propose a new CF-based recommendation methodology that constructs an ordinal scale-based customer profile under the implicit ratings condition. An experiment with the Web transaction data from a real online shopping mall shows that the proposed method performs better than existing CF methodologies. © 2013 International Information Institute.