The Research On Finding And Early Warning Model Of Internet Hot Public Opinion

In recent years, with the rapid development of Internet network in our country, network has been playing an important role in political and social life. Internet public opinion as a social extension of public opinion in the network, not only reflects the social reality of the various problems, but also would has a significant impact on the real world. Internet public opinion, especially the Internet hot public opinion, has features of widespread, dissemination of speed and high-impact, so it is necessary to research monitoring technology of Internet hot public opinion, also it is significant.The research of network public opinion at home and abroad mainly is in two areas: one is the theory research, including policy, mechanism, the user behavior characteristic, network transmission mode; the other is mainly network monitoring and warning system of public opinion, including web content analysis, the semantic grab, identification and data mining, theme identification and tracking. With the rapid development of domestic network, network public opinion studies increasingly become urgent, however, compared with current domestic and foreign studies, there is a big gap.This thesis introduces the hot public opinion research at present, and explains the basic knowledge of Chinese information processing, in the process of public opinion. In the model design, firstly, the characteristics of public opinion can be extracted from the hot public opinion expression. Based on the expression, processing has been proposed. And detail designs have been put on the content analysis module, the numerical analysis module and the cluster analysis module of the model. Finally, this thesis, based on the research of the current public opinion surveillance platform and practical demand, fulfils design and implementation on experimental platform, and carries out experiments on the model for effective verification.