点击次数：

论文类型：会议论文

发表时间：2018-01-01

收录刊物：CPCI-S

卷号：11168

页面范围：158-169

关键字：Sentiment word; Relative branch entropy; Sentiment analysis

摘要：There are a lot of new sentiment words appear in Weibo platform every day in the web2.0. As the unpredictable polarity of massive new words are detrimental to sentiment analysis for Weibo, hence how to extract new sentiment words and expand sentiment lexicon is of great importance. Therefore, we propose a relative branch entropy based method, which combines word frequency and adjacent words information to extract new sentiment word in Weibo. After integrated context and other factors, this method improves the accuracy of new sentiment word recognition. Three experiments are implemented on COAE 2014 Weibo corpus to compare the performance of different statistics with the proposed method. Experiment results show that the proposed method has a high accuracy, which demonstrates the effectiveness of this method and verify the promoted effect of new sentiment word extraction on the performance of Weibo sentiment classification.