Learning Hierarchical Review Graph Representations for Recommendation

The user review data have been demonstrated to be effective in solving different recommendation problems. Previous review-based recommendation methods usually employ sophisticated compositional models such as Recurrent Neural Networks (RNN) and Convolutional Neural Networks (CNN) to learn semantic representations from the review data for recommendation. However, these methods mainly capture the local and consecutive dependency between neighbouring words in a word window. Therefore, they may not be effective in capturing the long-term, non-consecutive, and global dependency between words coherently. In this paper, we propose a novel review-based recommendation model, named Review Graph Neural Network (RGNN). Specifically, RGNN builds a specific review graph for each individual user/item, where nodes represent the review words and edges describe the connection types between words. A type-aware graph attention mechanism is developed to learn semantic embeddings of words. Moreover, a personalized graph pooling operator is proposed to learn hierarchical representations of the review graph to form the semantic representation for each user/item. We compared RGNN with state-of-the-art review-based recommendation approaches on real datasets. The experimental results indicate that RGNN usually outperforms baseline methods in terms of mean square error.