The Research On The Crisis Warning Of Internet Public Opinions Based On Grey-theory

With the rapid development of Internet,network inforamtion and the amout of the people accessing the Internet is growing at an unprecedented rate. Internet is becoming an important way for people to obtain information and to express opinions and it is also the media to communicate with each other. With the performance of network hardware improving and prices of hardware and software decling,the internet users in China are also increasing. With such a large scale of Internet users and high popularity, Internet is becoming a huge space of public opinions with its charaters of freedom, real-time and rich in information. Without correct guidance, it can easily lead to network public opinion crisis,even the reality of mob behaviors that can casuse social instability.Internet public opinion reflects the sentiments of public opinion and it is also an important way to understand the reality of public opinion.So early warning of network public opinion crisis is very important for government decision-making to take appropriate measures to avoid the crises of Internet public opinion and guide reality work.The current researches of network public opinion are more in qualitative research, less in quantitative researches. Both the quantitative research and qualitative researches are mostly static description, less dynamic real-time warning.This paper first introduces the theoretical foundation and related concepts. Then,for the convenience of study, according to the law of the life cycle evolution,it presents the evolution of network public opinion by three phases: the initial propagation stage, the rapid spread of the stage and fading phase.Followed by a detailed analysis of characters of the evolution of each stage and that influence factors of Internet public opinion, it establish models to predict according to the dominant factors. The fouth chapter establishs models to give warnings by case study.In the first stage of the evolution of Internet public opinion, the main factors influence the evolution of the Internet public opinion is users'sensitivity to the internet public opinion or events. Therefore,in this statge, the concern of the users reflects the changes of network public opinion. So in this stage, it mainly establishes models according to the concern of the users that is a inial warning. In the second stage of the evolution of Internet public opinion, the dominant factor affecting the evolution of the Internet public opinion is the groups of users, the power of a collection of groups can promote the further development of the Internet public opinion. Groups of users is large in scale with a great influence on the range and it affects the spreading of the network public opinion greatly. The number of posts reflects evolution of the network of public opinion,so in the second stage, it mainly establishes models according to the number of posts.In the third stage of the evolution of Internet public opinion, Internet users become no longer as interested in the events as before,and they don't post as active as before. So the degree of the activity of the internet users reflects the change of the network of public opinion. In the third stage, it mainly establishes models according to degree of the activity of the internet users.The following models are referred to in this thesis:the gray GM(1,1) model, gray wave forecast, gray interval prediction, gray abnormal value prediction.