Application of Neural Graph Collaborative Filtering in Movie Recommendation System

With the gradual development of movie recommendation system, in order to improve the recommendation effect of movie recommendation system, we must learn how to get a more effective embedding. The main purpose of this paper is to combine the neural graph collaborative filtering with the movie recommendation system. It utilizes the user-item graph data by propagating embeddings on it. Owing to the method, the expressive modeling of high-order connectivity in user-item graph could injecting the collaborative signal into the embedding process in an clear way. We make a large number of experiments on MovieLens dataset, and verified the effectiveness and correctness of the algorithms we used.