News topic-typed microblog opinion sentence recognition

In recent years, with the development of the microblog platform, text mining technology in microblog is also developing fastly. Microblog's opinion sentence recognition task is the prerequisite of emotional polarity classification and public opinion analysis. Because of news topic-typed microblog's colloquial feature and the freedom of language structure, we put forward a combination classification method to identify the opinion sentence in microblog. Based on rules and AdaBoost, we use three filter rules and news topic-typed microblog's text features as weak classifiers of AdaBoost algorithm. Comparing with the traditional dictionary method and SVM method, the experimental results show that the method of this paper has achieved a good performance in the task of identifying the news topic-typed micro-blog's opinion sentence.