UBAR: User Behavior-Aware Recommendation with knowledge graph

The recommendation system is widely used in many aspects of digital economy to offer personalized services, in which efficient capture of user–item relations is of critical importance. However, there are two inevitable challenges in this task. On the one hand, the extraction of complicated associations is not easy among multiple users’ actions such as searching, browsing or purchasing. On the other hand, the integration of numerous items’ connections is indispensable for the recommendation framework. To address the stated challenges, we propose a User Behavior-Aware Recommendation method with knowledge graph (UBAR) consisting of a user behavior-aware module and an item knowledge graph module. The performance of the proposed UBAR method is evaluated on four datasets (i.e., Tmall, Taobao, Amazon, and Movie-Lens), and the experimental results demonstrate that the proposed UBAR outperforms state-of-the-art methods. The qualitative and quantitative analysis proves the effectiveness and efficiency of the proposed UBAR method.