Product feature and opinion word extraction is very important for fine granular sentiment analysis. In this paper, we leverage large-scale unlabeled data for jo
Data sparsity is a well-known challenge in recommender systems. Previous studies alleviate this problem by incorporating the information within the correspondin