A Novel Neural Collaborative Filtering Recommendation Based on Side Information Fusion

Abstract

It is difficult to accurately learn user's latent features using only one single data source. In order to solve these problems, we consider to utilize relevant side information of users or items as a supplement to rating information to enhance the performance of recommender systems, and propose a novel neural collaborative filtering recommendation model based on side information fusion. Extensive experiments on different datasets validate the efficiency and accuracy of our proposed framework.