Chinese text sentiment classification based on granule network

Abstract

With the expanding of text comment information, text sentiment classification become a hot issue. Domestic research on chinese sentiment classification mainly focus on segmentation and features selection or focus on classifying algorithm based on statistics. Rules mining method is a kind of important techniques of text classification. This paper propose a new approach which apply the rule mining by granule network constructing to sentiment classification. Approved by experiment result, text sentiment classification based on granule network is effective.