Sentiment Classification of Stock Comments in Chinese Based on Semi-Supervised Approach Correcting with Feedback information

摘要

This paper proposes a sentiment classification approach of stock comments in Chinese, which is a new research field of text polarity analysis. Method in this paper firstly uses a semi-supervised text categorization approach based on bootstrapping to solve the classification; Secondly, because not every portion of a document has the same polarity with the overall sentiment of the document, so a correcting step with feedback information is added into this approach. The results show that statistical method can work well without lexicon on sentiment classification of stock comments in Chinese, and the feedback information can improve the result to a certain extent.