Sentiment Analysis Of Chinese Microblog And Its Application

In recent years, the rise of microblog has gradually become a mainstream news release channel.A large number of users use microblog to pass information, put ideas or express feelings rapidly andtimely. Hot events and topics in net can be also spread quickly and rapidly, these public views canbe collected through the microblog. For these reasons, in recent years, the study of the sentimentanalysis in microblog has become a new hotspot.In the sentiment analysis, there are three main emotional tendencies: positive tendencies, neutraland negative tendencies. Traditional sentiment analysis methods include machine learning basedand emotion dictionary based approaches. This thesis takes the second method to analyze Chinesemicroblog sentiment.The main work in this paper include:For the characteristics of Chinese microblogs, we used a Chinese word segmentation based onthe combination of several traditional method to pretreat Chinese microblogs. This method cansolve the crossing ambiguities that often appear in the segmentation process and extract usefulemotion elements from texts. Then the useful emotional elements in microblog, such as emotionalwords, adverbs of degree, negative words, emoticons.etc are analyzed. Finally, the wholemicroblog emotional polarity and emotional intensity derive from the results of the analysis of allemotional elements..Considering that network irony phenomena exists in Chinese microblogs, based on the emotionpolarity and emotion intensity, this paper propose a sentiment analysis method based on MMTD inorder to get a final emotion tendency estimate for each microblog. This paper selected8hot topicsin SINA microblog website related to different social hot spots, and used the proposed sentimentanalysis method based on MMTD. The experimental results show that the method can find the ironyphenomena exist in Chinese microblogs, that helps to improve the accuracy of the microblogemotion tendency estimatings.