Chinese Text Sentiment Analysis Based on Extended Sentiment Dictionary

• Main Authors: Guixian Xu, Ziheng Yu, Haishen Yao, Fan Li, Yueting Meng, Xu Wu
• Format: Article
• Language: English
• Published: IEEE 2019-01-01
• Series: IEEE Access
• Subjects: Chinese text sentiment analysis; text classification; naive Bayesian; sentiment dictionary
• Online Access: https://ieeexplore.ieee.org/document/8675276/

The method of text sentiment analysis based on sentiment dictionary often has the problems that the sentiment dictionary doesn't contain enough sentiment words or omits some field sentiment words. In addition, due to the existence of some polysemic sentiment words with positivity, negativity, and neutrality, the words' polarity cannot be accurately expressed, so the accuracy of text sentiment analysis is reduced to some extent. In this paper, an extended sentiment dictionary is constructed. The extended sentiment dictionary contains the basic sentiment words, the field sentiment words, and the polysemic sentiment words, which improves the accuracy of sentiment analysis. The naive Bayesian classifier is used to determine the field of the text in which the polysemic sentiment word is. Thus, the sentiment value of the polysemic sentiment word in the field is obtained. By utilizing the extended sentiment dictionary and the designed sentiment score rules, the sentiment of the text is achieved. The experimental results prove that the proposed sentiment analysis method based on extended sentiment dictionary has certain feasibility and accuracy. The research is meaningful for the sentiment recognition of the comment texts.