Joint Attention LSTM Network for Aspect-Level Sentiment Analysis

Abstract

Aspect-level sentiment analysis, as an important type of sentiment analysis, is a fine-grained sentiment analysis task which has received much attention recently. Recent work combines attention mechanisms with neural networks to learn aspects feature and achieves state-of-the-art performance. However, the prior work ignores the sentiment terms feature and the latent correlation between aspect terms and sentiment terms. In order to make use of aspects terms and sentiment terms information, a method that based on joint attention LSTM network (JAT-LSTM) for aspect-level sentiment analysis is proposed, which aspect attention and sentiment attention are combined to construct a joint attention LSTM network. The experimental results on the benchmark datasets show that the proposed method achieves better performance than the current state-of-the-art.