摘要
1 Introduction Recommender systems can effectively alleviate the problem of information overload.However,traditional recommendation methods cannot capture users’dynamic interests.Sequential recommendation methods model user sequences to obtain more accurate and dynamic user interests.Recently,deep learning-based sequential recommendation methods have achieved great success.RNN is proposed to capture the sequential information[1,2].Attention-based methods[3]use attention mechanisms to learn relationships between items.GNN-based methods[4−6]transform sequences into graph structures to capture relationships of items.However,they have the following two limitations.
基金
the National Natural Science Foundation of China(Grant Nos.62172283 and 62272315).
