Sentiment classification via user and product interactive modeling

Xiabing Zhou, Zhongqing Wang, Min Zhou, Qifa Wang, Shoushan Li, Min Zhang, Guodong Zhou

Abstract: Sentiment classification aims to identify the polarity of a given review. Most existing methods consider each review as an individual while ignoring the importance of the user and product information of the given review. A direct way to integrate user and product information is to employ an attention mechanism to learn the local interaction between them. However, local interactions cannot capture the global optimization among user and product information. Therefore, we propose a novel interactive model to integrate both local and global interactions between users and products. In particular, we employ an attention mechanism to learn local interactions between users and products, and construct user and product interactive graphs to model the global interaction of users and products. Empirical evaluation shows that our model outperforms previous state-of-the-art methods significantly by learning the local and global interactions among users’ preferences, product characteristics, and reviews.