Sentiment Analysis of Chinese Online Reviews Based on Word2vec and DBN

Abstract

Sentiment Analysis of Online Reviews has a broad application prospect in the commercial field as it truly reflects the user satisfaction with the product. However, the online reviews belong to short text and the feature words are sparse. The traditional classification algorithm makes a very little effect. And the traditional method based on the bag-of-words model always ignore the relationship between words and lose the context information, resulting in poor classification results. In this paper, we have proposed a method combining Word2vec and DBN towards the sentiment classification for Chinese online reviews. Experimental results show that our method has good performance.