A Personalized Recommender Integrating Item-Based and User-Based Collaborative Filtering

Recommender systems employ prediction algorithms to provide users with items that match their interests. The collaborative filtering (CF) is the most popular system and the two of the most famous techniques in CF are the user-based CF (UBCF) and item-based CF (IBCF). Nevertheless each of them takes only one-directional information from the user-item ratings matrix to generate recommendations. In other words, the UBCF utilizes user similarities and the IBCF tries to make a prediction by utilizing item similarities. It means that methods may use only half of the total information from the given data set. For completing the missing part of usable information, this paper proposes a CF algorithm integrating the UBCF and IBCF, which takes both vertical and horizontal information in the user-item matrix. It produces perdition using IBCF to form a dense user-item matrix and then recommends using UBCF based on the dense matrix. The experimental results on MovieLens dataset show that the proposed algorithm outperformed in terms of prediction accuracy and robustness to data sparseness.