Collaborative Filtering Based on Attention Mechanism

Each item has different characteristics. Different users place different importance to these aspects which can be thought of as a preference/attention weight vector. Therefore, using the traditional method to predict the unknown ratings ignores this difference in importance, resulting in a false assumption that all users have the same attention to different characteristics of the same item. In this paper, we propose a collaborative filtering system based on attention mechanism and design the feature-topic model to extract the characteristics of the item from review texts. Then, we use an attention network to get the importance of the item’s characteristic to the user. Considering the shortcomings of linear interaction features, we adopt the idea of collaboration to predict unknown scores. Extensive experiments on real-world datasets show significant improvements in our proposed CFAM framework over the state-of-the-art methods.