The Study Of Network Public Opinion Early Warning On Emotional Computing

In recent years, with the rapid development of science and technology, the network is playing a significant role in our country's political and social life. And the it has become an important platform for people to express their ideas and emotions,given its virtuality and anonymity. With the increase of social emergencies, a linkage online-offline situation has been formed. Doing nothing but letting their feelings spread indulgently will easily result in network public opinion crisis, even big riots.Hence, how to find out and prevent it turns a critical issue to be solved.The study of network public opinion nowadays at home and abroad mainly concentrate on two aspects : one is theory research, including the policy and mechanism, user behavior characteristics, and the network public opinion propagation mode; The other one is the study of the network public opinion which is about the monitoring and warning system, such as the web scraping, content analysis,semantic recognition, data mining, themes recognizing and tracking algorithms and techniques. However the study in the view of emotional computing lacks.After reading much literature on emotional computing and network public opinion crisis, the paper uses the methods of data mining and emotion assorting when doing its exploratory study on the basis of network texts. The research framework of this paper mainly consists of four components, namely theoretical background, emotional index classification, the design of early warning model on emotion computing and the empirical example of early warning model.The paper is not only somehow in the leading edge of the study on emotional index design, the information acquiring of network texts and early warning model,but also has great practical instructional value on the proper monitoring of network network public opinion.