Add to Chrome
Add to Firefox
Add to EdgeSession-based Recommendation with Graph Neural Networks
Nov 05, 2018
The problem of session-based recommendation aims to predict users' actions based on anonymous sessions. Previous methods on the session-based recommendation most model a session as a sequence and capture users' preference to make recommendations. Though achieved promising results, they fail to consider the complex items transitions among all session sequences, and are insufficient to obtain accurate users' preference in the session. To better capture the structure of the user-click sessions and take complex transitions of items into account, we propose a novel method, i.e. Session-based Recommendation with Graph Neural Networks, SR-GNN for brevity. In the proposed method, session sequences are aggregated together and modeled as graph-structure data. Based on this graph, GNN can capture complex transitions of items, which are difficult to be revealed by the conventional sequential methods. Each session is then represented as the composition of the global preference and current interests of the session using an attention network. Extensive experiments conducted on two real datasets show that SR-GNN evidently outperforms the state-of-the-art session-based recommendation methods and always obtain stable performance with different connection schemes, session representations, and session lengths.
