Public Sentiment Classification And Reason Analysis For Micro-blog Events

Wit h the rap id development of socia l media, especia lly micro- blog, more and more people are getting used to the Internet to express the ir views on hot spots. How to make full use of these mass ive data about the hot events, to the hot spots and the public opinion to carry on the thorough ana lys is, has become a very meaningful topic. This article has done the thorough research to this topic.Sent iment class ification of micro-blog data on hot events has become an important tool for researchers to analyze hot events and public opinion. The research o f this paper is about the micro- blog e motio n class ification. A nd in this paper, we have carried out in- depth research work fro m two aspects : the promot ion of micro- blog sent ime nt classificat ion effect and the cause ana lys is of micro- blog event sentiment distribution.In order to improve the micro-blo g sent iment c lassificatio n effect, this paper puts forward the improved scheme in two aspects, and the related experiments are carried out. The y are: micro-blog sent iment classificat ion confus ion posit ion category micro- blog and micro-blo g w ith e mot iona l proble ms, as well as in t he emotio nal class ification of micro-blo g did not cons ider the user’s histor ical informat ion can be improved. O n the first question, this paper adopts the method of adding t he opinio n category and the micro- blog c lassificatio n algorit hm based on sent iment class ification. For the latter, by add ing the user history micro- blog informat ion to the exist ing sent iment classificat ion algor it hm, the feasib ility of adding user information is verified through experimental comparison.About the cause analys is of micro-blog event sent iment distr ibut ion, this paper also studies and ana lyzes fro m two aspects: the emotiona l distr ibut ion and the emotio nal change point. First, this paper defines the reason analys is, in essence, the process of extract ing t he most releva nt reasons for the event of micro- blog events. Because the process of the extraction of sub events is required in the eve nt of s ub event c luster ing, and in a number of sub eve nts to extract the most re levant cause of the sub event is required to document the sort. So in this paper, we carried out the sub event clustering, the reasons for the sub event sort of the two parts of the research and experime nt, all get good results.