Search

The session-based recommendation predicts the next user's interest item based on an anonymous user–item interaction sequence. However, most existing methods focus on capturing sequential signals or item-transition patterns within the current session while ignoring potential collaborative behaviors among different users from other sessions that could positively affect the recommended performance of the current session. To address these issues, we propose a Neighbor-Enhanced Graph Transition Network , which uses a diverse graph neural network to model complex interactions at the item level between the current session and its neighboring sessions. We create a Current Feature Encoder to investigate the user's current preference and a Neighbor Feature Encoder to generate useful collaborative information by considering the popularity of item-transition pairs from neighbor sessions. Then, we propose a fusion function that combines the two types of features mentioned above. We use a positional attention mechanism to investigate the impact of items in different positions on the user's true intention. The experimental results over three real-world datasets demonstrate that our proposed model generally outperforms other state-of-the-art methods. [ABSTRACT FROM AUTHOR]