Sentiment Classification of Chinese Text Based on Extending Semantic Similar Sentiment Words

Abstract

Aiming at the problem of sparse sentiment features caused by the lack of effective sentiment words in Chinese text sentiment classification, a high-precision network comment sentiment classification method is proposed to expand semantic similarity sentiment features. Experiments based on takeout and computer review corpus show that the accuracy of sentiment classification is high and the classification effect is good. Automatic sentiment analysis technology was used to mine the sentiment tendencies contained in a large number of comment texts. It can understand the opinions of the public view on an event or product, and provide decision support for making marketing strategies and selecting commodity brands. The method has important commercial