Chinese review sentiment analysis based on deep learning and attention mechanism

Abstract

With the development of the Internet and the continuous growth of the number of netizens, more and more people publish their attitudes and emotions on products, services, etc. on the Internet, accumulating a large number of comments containing personal opinions. The sentiment analysis of text reviews is of great significance to explore the emotional tendencies of users and help merchants adjust product positioning and marketing strategies in a timely manner. The deep learning framework adopts CNN and LSTM models respectively to introduce the attention mechanism for sentiment classification. The effects of CNN model and LSTM model on sentiment classification in Weibo comments were analyzed. From the experimental data, after the same number of iterations, the accuracy of the CNN model reaches 97.94%, and the accuracy of the LSTM model reaches 98.18%. The experimental results show that LSTM is better than CNN in the sentiment classification of Weibo comments.