Research And Application Of Network Public Opinion Evolution Model

Network public opinion is a product of the impact of globalization and informatization.It's a collection of views,opinions,attitudes and emotional information expressed by people in informatizational way.Network is not only a kind of social supervision,but also a direct reflection of social public opinions.Therefore,with the rapid development of Internet technology in the global scope,the impact of network public opinion on political life order,economic development and social stability is increasing.Hence,more and more domestic and foreign scholars pay attention to the network public opinion.Based on this,the research on the evolution rules,characteristics and influencing factors of network public opinion,prediction and analysis of network public opinion dissemination trend and impact,help the government management and control of public opinion and development trend of enterprises to deal with emergencies,dealing with the network public opinion,to avoid the negative impact of the expansion.In this paper,we establish the evolution model of the network public opinion and research on the prediction of network public opinion based on the analysis of the evolution mechanism and dynamic characteristics of the network public opinion.The main research contents are as follows:(1)Analysis and research of network public opinion's concept,concept of evolution,main body,dynamic evolution mechanism and the basic model.And then determine the basic characteristics and influencing factors of network public opinion.All of those are preparing for the establishment of network public opinion evolution model.(2)Cellular automaton can constitute the evolution of complex dynamic systems through simple local interactions.This paper establishs the network public opinion evolution model based on cellular automaton according to using the theory of system dynamics,combining with the characteristics of the network public opinion,making the network public opinion evolution rules.Based on the ideology of the interaction between online and offline line network,the double coupled cellular automaton evolution model is composed,which is through the interaction of evolution rules.The two cellualar layer grids are consisted of online and offline network.Then we analyse the influence of different levels and factors on the process of network public opinion information dissemination.(3)The network structure of the cellular automaton is a regular network.But the actual link between the users is a complex network.Combined with complex network theory and network public opinion evolution rules,the network public opinion evolution model are established in two different network topologies,including NW small world network model and BA scale-free network model.This paper analyzes and studies the evolution law of network public opinion under different network topology and different evolution rules.(4)Research on forecasting the time series of network public opinion.First of all,the trend index of key words collected on the Baidu index platform is used as the time series of the network public opinion.Then we get the reconstruction vector using the the phase space reconstruction of chaotic theory.Ultimately,we judge whether the network public opinion time series possesses chaotic characteristics via the maximum Lyapunov index.The simulation results show that chaotic characteristics exist in the network public opinion development time series.The forecasting model based on echo state network is put forward.Then we improve the model using the complex network instead of reserve pool,solving the output weights with two norm regularization model,using hybrid co-evolutionary algorithm to optimize the structural parameters of prediction model.The improved prediction model is used to predict the network public opinion.Some chaotic time series are predicted by this way in order to verify the validity of the improved prediction model.Through the simulation experiments of the network public opinion and chaotic time series,the results show that the improved model has the advantages of accuracy,rapidity and adaptability,and can also be used in other fields.The evolution model and prediction model of network public opinion are proposed,which provides a new way for the study of network public opinion,enriches the theory and method of network public opinion evolution,to provide theoretical support and reference value for the guidance and control of the network of public opinion,has important significance in maintaining social stability and promoting development.