Agent-based User Behavior Modeling In Social Network And Experimental Analysis

Social network user behavior modeling aims to maximize the simulation of human real behavior and analyze the law of it,and then give some reasonable explanation of behavioral phenomenon.However,early user models only treat the user as a node in a social networks,And the user influence models base on graph theory or the behavioral dynamics models based on statistics haven't explore the internal mechanism of behavior.While some models involve internal decision-making mechanisms,they are not comprehensive enough,so it is difficult to achieve personification decision.Based on the microscopic individual,a decision model of extended closed-loop mental process is built in this thesis,and it is applied to the abstract social class network.The agent-based modeling method is used to construct the state transfer model.In this thesis,the basic framework of decision model is put forward based on the application scenario,and on this basis,the mental process of decision making is refined,four process models are constructed,including Emotional incentive model,Dual channel decision model,Intention volatility model and Experience feedback model.Then,user state transfer model of the abstract social class network is proposed in this thesis,in order to study on the social class mobility from the perspective of decision making.Finally,the superiority of the mental process model is verified in this thesis based on the data of self-driving traveling community.And the experiments based on Netlogo software verify the validity and correctness of the state transfer model.According to the experiments,some conclusions are as follows: It is possible to improve the prediction accuracy of the model by constructing a sound mental process,and in the decision making situation with longer interval,users may change their original intentions.In addition,the state transfer model proposed in this thesis is fitted to the actual class distribution curve.And another conclusion is that the greater the user's internal motivation is,the easier it is for the user to take state transfer.