Research On The Dissemination And Prediction Model Of Public Opinion Forwarding Volume On Emergency Network

Emergencies are closely related to people's lives and emerge in an endless stream.The public opinion caused by emergencies is widely concerned and has a negative impact on social stability.The rapid development of network technology provides a good support for information dissemination.With the help of network,public opinion in emergencies is widely spread,which brings new challenges and difficulties to effective supervision and monitoring and benign guidance.Therefore,the network public opinion of emergencies has become the focus of public security and emergency management.It has important theoretical significance and practical value for the relevant departments to study the dissemination mechanism of public opinion of emergencies in depth and build an efficient and reasonable prediction model of public opinion transmission volume by using advanced scientific methods.Firstly,this paper combs the research results and shortcomings of the evolution law,the influencing factors of public opinion and the prediction model of network public opinion.Then we expound the related concepts and theories of emergency and network public opinion.Taking food safety events as an example,the selection of research target events and their processing methods are introduced to provide data support for subsequent research.Through the introduction of social network and complex network theory,taking " Gang Rong steamed cake unqualified " event as an example,we built emergency public opinion communication network.Then we analyzed the rules and characteristics of public opinion communication network,and further extracted the relevant factors that affect the public opinion communication network.Combining with the existing literature research,from the user influence and micro blog content influence,a more perfect and reasonable prediction index system of individual information communication of network public opinion of emergency was constructed.Then we introduced the inertial weight differential descent method to improve particle swarm optimization(Improved particle swarm optimization,IPSO)and constructed a network based on IPSO-BP neural network prediction model of individual microblog forwarding volume.Finally,based on the model,we further explored the importance of quantitative analysis of each impact index,and based on the prediction model and case analysis,the effectiveness and accuracy of individual micro blog forwarding quantity prediction are applied and verified.The results show that the network of emergency public opinion is of loose structure,high accessibility,high local density and scale-free,which are important factors for the rapid spread of public opinion information.From incubation period to maturity period,the scale of public opinion network increased significantly,and opinion leaders and community structure became prominent.Opinion leaders lead and control the development of public opinion all the time.After maturity,the network of public opinion tends to be stable,except the scale increases,there is no obvious change.The user activity and fan activity of the key nodes in the network can change the location attribute characteristics of the nodes and affect the structure of the network;the relevant indicators of user influence and microblog content influence have an important influence on the forwarding and dissemination of public opinion information on the network,and the importance of the two indicators is not much different,at the same time,they are relatively speaking.The six sub factors,such as super topic,user activity and fan activity,are of high importance and have stronger influence on the forwarding and transmission of public opinion information compared with other sub factors.The IPSO-BP neural network prediction model for individual public opinion forwarding volume and transmission proposed in this paper has better advantages in accuracy and effectiveness compared with other models,which can be quickly and conveniently realized in public opinion events.The prediction of a large number of individual microblogs provides a basis for the effective research and judgment of the overall network public opinion development of emergencies.The research of this paper is helpful to improve the ability of public opinion monitoring and early warning,and provide theoretical and technical support for scientific emergency response.