Joint Neural Collaborative Filtering with Basic Side Information

Abstract

Nowadays, deep neural network has been greatly developed and widely used in many areas. However, the research of the deep neural network on recommendation system is inadequate. Most research focuses on analyzing the textual descriptions of items and comments of users, making use of the neural network to get feature vectors from texts or pictures. In this paper, we directly adopt the deep neural network to better fit the non-linear relationship of users and items and effectively integrate some side information (basic information and statistical information) into the neural network. Utilizing deep neural network, we explore the impact of some basic information on neural collaborative filtering. To the best of our knowledge, it is the first time to combine the basic information, statistical information and rating matrix by the deep neural network. Finally, we use the benchmark data set (MovieLens) to demonstrate the effectiveness of the proposed deep neural network model with side information.