Text sentiment classification based on the automatic expansion of sentiment dictionary

Abstract

The sentiment dictionary-based text sentiment classification method is one of the main methods in the field of sentiment analysis, and the completeness of the sentiment dictionary is one of the key factors of the method. With the constant appearance of the new terms on Internet, the static sentiment dictionary cannot cover all the sentiment words. In this paper the automatic expansion methods of sentiment dictionary is explored, and a method of automatic recognition and orientation of sentiment words based on reverse back stepping is proposed, and the recognized sentiment words are used to expand the static sentiment dictionary. By applying the expanded sentiment dictionary to sentiment classification, we found that the expanded dictionary can greatly improve the recall rate of the sentiment classification without reducing the accuracy rate, so as to enhance the overall sentiment classification performance.