CMF: Coupled Matrix Factorization for Recommender Systems

The challenges in Recommender System (RS) mainly involve cold start and sparsity problems. The essence behind these problems is that the extant techniques normally mainly rely on the user-item rating matrix, which sometimes is not informative enough for predicting recommendations. To solve these challenges, many existing papers considered user/item attributes as complementary information with an assumption that attributes are independently. In real world, however, user/item attributes are more or less interacted and coupled via explicit or implicit relationships. Limited research has been conducted for analysing and applying such attribute interactions into RS. Therefore, in this paper we propose a novel generic Coupled Matrix Factorization (CMF) framework by incorporating the coupling relations within users and items. Such couplings integrate the intra-coupled interaction within an attribute and inter-coupled interaction among different attributes to form a coupled representation for users and items. Experimental results on two open data sets demonstrate that the user/item couplings can be effectively applied in RS and CMF outperforms the benchmark methods.