Micro-blog Emergency Network Public Opinion Research On Rule Discovery And Prediction Method

The arisen of emergency events and the popularization of the network makes emergency network public opinion come more frequently and cause more harmfulness. Due to the cognitive surplus and technical skills, Micro-blog has become an important channel for the spread of emergency network public opinion. Therefore, using of the existing means of information technology to predict emergency network public opinion is of great significance for emergency decision making and the maintenance of social stability.Though the research on emergency network public opinion prediction has obtained certain achievements, there still exists many shortcomings. On the one hand, most researches on emergency network public opinion prediction method focus on how to improve the precision of the model or method, which lacks of considering the emergency network public opinion characteristics. Therefore, at the same time of improving the precision of the model, we also need to study the internal rules and characteristics of the network public opinion system, so as to provide a better theoretical support for the network public opinion prediction research. On the other hand, most researches on emergency network public opinion rules identification carried on from a macroscopic angle, such as the evolution characteristics and the propagation. This leads the lack of the relevant rules study. In addition, rules study' guiding role towards public opinion monitoring and forecasting is not fully embodied in the process of practical application.In order to solve the problem of the existing research, we firstly recognize and analysis microblog emergency network public opinion influence rule between emotions and behavior through several methods, such as text analysis, emotional calculation and Multilevel Time-Series Regression Model. Put forward the corresponding suggestions based on the rule. Then identify the main factors that cause the Micro-blog number fluctuations through unit root test and granger causality test method, and then explain the reason with related theory. Then Build Micro-blog emergency network public opinion prediction model based on the grey prediction model, and optimize the model utilizing emotional characteristics of the Micro-blog emergency network public opinion. Finally, by gathering a large number of typical Internet public opinion emergency Micro-blog data, we predict and verify the validity of the model.Begin with the study of Micro-blog network public opinion internal elements, this paper explore the influence law between Micro-blog behavior, mood and the internal factors that affecting the number of Micro-blog fluctuations. This provides a theoretical basis for other emergency network public opinion research. On this basis, then build prediction model and carry on the optimization, which enhances the precision of forecasting model and make it more suitable for forecast Micro-blog emergency network public opinion. Besides, it also provides a technical support for emergency network public opinion in the forecast.