Research And Implementation Of Microblogging Network Measurement And Topic Trend Forecasting Model Of Emergencies

When an emergency occurs, usually related topics will spread on the network, such as through the microblogging platform. Because of its suddenly, uncertainty, limited information of emergencies, errors, partial and radical opinion would cause people to panic behavior, and undermines social stability brought to the majority of the people’s material and spiritual losses. So microblogging network measurement and topic trend forecasting is very important in terms of public opinion analysis and early warning in network.By studying the measurement of the Sina microblogging network under emergency, this page is aim to analyze the characteristics of microblogging topic formation, dissemination and development.Network measurement is composed by crawling and analysising data of Sina microblogging--our country’s the largest microblogging platform, measuring indicators of microblogging network topology. By describing the topology Sina microblogging network under emergency, this page analyze the reasons why emergency topic on Sina Weibo spread rapidly, and establish a topic trend prediction model after an emergency occurred on this basis, finally design and implement a prototype system about topic trend forecasting, which provides a reference for the study of issues related to monitoring public opinion.In this paper, the work is divided into two part, the first is the measurement of microblogging network under emergencies, the second is the prediction of topic tendency under emergencies. First, use self-developed Java crawler to get microblog data which is associate with emergencies from Sina Weibo. Secondly, through data preprocessing, parameter measurement and compute, experimenting and result analyzing to complete the work of network measurement. Finally, through topic discovery, establishing model and evaluating, experimenting, result analyzing to complete the researching and implementation work of topic tendency prediction model.This paper compared the result of microblogging network measurements under emergency with existing online social network measurement results, found that the microblogging network is a small-world network, its characteristic path is short, degree of polymerization is high, and customer relationship in network is close, users tend to connect the node with high degree. These characteristics explain why emergency topic in Sina Weibo spread so rapidly, and there are opinion leaders in the microblogging network under emergency. On the basis of network measurements, this paper selected hidden Markov model(HMM), combined with the state of opinion leaders participating the discussion of emergency, predicted the topic trend of emergency, established the corresponding model, and designed and implemented a microblogging network measurement and topic prediction prototype system, using actual data to evaluate the experimental results, results showed the error is small.