Leveraging Explicit Products Relationships for Improved Collaborative Filtering Recommendation Algorithm

Abstract

Collaborative filtering (CF) is widely applied in recommender systems. This study aims to incorporate product relationships into traditional CF based recommendation algorithms to address data sparsity and cold start problems. We propose a novel matrix factorization model, where each user has a category-specific user latent factor vector to more accurately describe the user latent factors. In the meanwhile, the explicit product relationships have been leveraged as the regular term to constrain the learning of product latent feature vector. Lastly, the empirical results demonstrate the superiority of our model against other counterparts on recommendation accuracy. Besides, our model has a good theoretical and practical significance.