Integrating feature information for improving accuracy of collaborative filtering

Abstract

Collaborative filtering (CF) has been widely used and successfully applied to recommend items in practical applications. However, the collaborative filtering has two inherent problems: data sparseness and the cold-start problems. In this paper, we propose a method of integrating additional feature information of users and items into CF to overcome those difficulties and improve the accuracy of recommendation. We apply a two-pass method, first filling in unknown preference values, then generating the top-N recommendations. © Springer-Verlag Berlin Heidelberg 2004.