Simulation Of Internet Public Opinion Orientated Agent Modeling Of Web Users

A person,as an automatic individual,plays an important role in the occurrence and development of emergency.Groups composed by some individuals act as bearing body against emergency.If we want to do research of what role individual plays in the occurrence and development of the emergency,it is very important to take care of modeling of individual behaviors.This paper,orientating the simulation of Internet public opinion,analyzed the process of information in psychology and behavior,studied the theory and method of association and integration of multi-source data,proposed reduced-order model based on complete mechanism model,finally,the individual model is implemented and the parameters are analyzed based on the actual data.The main research work and achievements are as follows:(1)A complete mechanism model of the psychological behavior process of individual processing information is established.After several discussions with domain experts,this paper analyzed in detail the whole process of online information processing,including information receiving,information cognition and information dissemination,and analyzed qualitatively the factors that need to be considered in each process and the mechanism that factors worked,and finally got the mechanism model,which reflected the complete psychology behavior of online information processing.(2)Based on the complete model,a reduced-order model based on network data is established.The complete model has the characteristic that the factor and the factor function process can only be qualitatively descripted,but cannot be quantitatively described.Therefore,in this paper,under the guidance of domain experts,the model was reduced order,the parameters of the model can be extracted from the data,at the same time in line with the psychological process of information processing.Finally,the realization process of the model was expounded.(3)The methods and techniques of multi-source network data association and integration are studied.This paper studied the methods of data acquisition and association and integration,and proposed a multi-source data association integration framework suitable for individual character modeling in this paper.Based on the actual network data,we do data analysis and correlation and integration.(4)For the realization of the reduced-order model,the individual and network attributes are analyzed,and the method and results of parameter assignment are studied.The experiment is constructed and the model is validated.Aiming at the realization process of the model,the network attribute,the individual attribute,the individual media literacy analysis experiment based on the real network data,and the experiment of the parameter assignment in the model realization are carried out.The problem of parameter assignment and the method of further solving are also found.Finally,in the first half of 2012,part of the micro-blog users' data for the training set to July data for the prediction set,construct the experiment to test the model predictive effect.Finally,the conclusion that the model can basically meet the demand is obtained.The complete mechanism model fully reflects the psychological behavior process of the individual processing information from the theoretical mechanism level,and has done the theoretical level design for the individual character modeling of the network public opinion,which can guide the individual modeling of related fields.The reduction model based on the network data sets up a bridge for the mechanism model and the engineering realization of the model,and establishes the foundation for the engineering application.