To solve the problem that existing session recommendation methods take less account of the conversion information between items from other different sessions,a session recommendation method based on global enhanced graph neural network was proposed.The position-aware global graph was built to distinguish the different types of neighbors by utilizing the relative position information,to model the neighboring item transition to train global-level item embeddings.Local-level item embeddings were trained using attention mechanisms to capture transition information from the current session.The inter-session embeddings and intra-session embeddings were generated from two types of item embeddings.The contrastive learning technique was used to enhance the robustness of two types of session embeddings.Experiments on three datasets demonstrate the effectiveness of the method.
维普
万方数据