The Research Of Emotion Analysis Techniques For Chinese Microblog

In recent years, micro-blog is a network media which has a fasting development. As having real-time, openness, interactivity and timeliness, casual style and so on, more and more people are using micro-blog. And then it affects our daily-life, so there is a strong social powers and communication. There is no doubt that a wealth of worthy information are hidden within micro-blog. It can provide many info to government, company and person by analyzing the micro-blog used the technology of emotion analysis. Fine-grained sentiment analysis become the main study in sentiment analysis as the research developing. By analyzing the micro-blog’s text content including expression and other related content, this paper would find out the micro-blog belong to which kind of emotion.The main research content of this paper include the following:First of all, analyzing the characters of language performance about micro-blog. Comparing with the general text, micro-blog which belong to internet phrases has many new characters in language performance and this paper would analyzing them in the point of emotion including symbol,number and expression.Then introducing the method of computing emotion words based on dependency parsing in detail. This section main include extracting relationship of emotion words by the analysis of dependency parsing. And four formulas formula of how to compute strength and polarity of emotion vocabulary are proposed.At last, the content is classified strategies based on SVM and language regulations. Although the method which using SVM could get a better result of classification then other methods, there are many deficiency about the result. So in order to get a better result, the language regulations which have twenty items included three negative regulations would be used to amend the result of classification used SVM.