Artificial Society Based Modeling And Simulation For Public Opinion Event

Recently,with the development of economy and society,a variety of institutional and structural contractions become increasingly prominent,and the differentiation and conflict of interests between social groups are growing.Mass contingency events caused by opinion spreading and emotion gathering appear to be on the rise.Through certain methods and techniques the features,mechanism,and laws of public opinion can be found.So the research on the evolution and management of public opinion has important theoretical and practical significances to construct a positive environment for the harmonious development of society.Considering the response and management of public opinion event,the paper summarizes the main idea to carry on the successive research.Then we propose a method of artificial society to achieve the quantitative research on the evolution and formation of public opinion.The key techniques are deeply discussed,including the evolution mechanism of public opinion event,the generation of social networks,human behavioral modeling,and standardized modeling.The main work and innovations of the paper are listed as follows:(1)The evolution mechanism of public opinion events is formally proposed.The evolution of public opinion event is defined as a dynamic process of social system in terms of structure,behaviors,and states over time.The time-varied process of the event is decoupled as a series of state transitions in the state space.Through defining the social structure,attributes of population,and human behaviors formally,the interactive relations among event,environment,and population are explained,and the evolution mechanism of public opinion is illustrated.(2)A model of opinion dynamics based on information exchange is designed.Traditionally,the researches on public opinion almost consider that the opportunity of interpersonal communication is equal,and neglect the process of information dissemination.In fact,the information acquisition is a premise of opinion expression and interaction.In order to fully reflect the multiple stages of the event happening,information dissemination related to event,opinion expression and exchange,and the formation of public opinion,the paper proposes an opinion dynamics model based on information exchange through the effective integration spreading dynamics and opinion dynamics.(3)A generation algorithm of social networks for artificial society is proposed.Considering the similarity and spatial distance among population,we assume that people always intend to gain utility from interacting with others,and the tie formation between any pair of individuals can be simulated as a social encounter.Based on human behavior choice theory,an evaluation and judgment mechanism is designed for generating social networks.The introduction of behavior choice theory can make it reflect the diversity and randomness characteristics of human behaviors.The algorithm is applied to generate egocentric networks for a synthetic population of interest,the results show that the generated social networks has the key features,such as small-world property,high clustering coefficient,and community structure,which comply with the empirical research.It can be concluded that the method is effective to some extent.(4)An algorithm for generating human daily activity is proposed.Based on need theory and discrete choice model,the design ideas of the algorithm are summarized as follows: need is the internal power of activity execution,and conducting an activity can meet some needs,and gain utility.The timing and duration of activity are simultaneously decided through a modeling framework of integrating individual and joint decision-making together.The algorithm can explain the motivation and internal drive mechanism of activity execution in the theory level,and generate daily activity schedule of arbitrary time length for each individual of synthetic population.Through a case study,we find that the algorithm can reflect some daily activity patterns(e.g.periodic pattern),which illustrates the effectiveness of the algorithm.(5)A standardized modeling paradigm of artificial society for public opinion event is also proposed.Based on the element,object,relationship,and process involved in the modeling of public opinion event,the standardization process of model development is given.We divide the objects into three big parts,including population,event,and environment,and design a meta-modeling framework for public opinion event.The meta-modeling framework can support the development of modeling language and tools,and help the domain-expert to design simulation model and implement an experiment of interest in the future.Finally,based on the previous works on model design,data generation,and simulation modeling,we carry on a case study to investigate the evolution process of public opinion by interpersonal contact in a large-scale artificial society.Through the case study,the practical significance of the research is demonstrated,and the advantages of artificial society in describing complex social systems are also discussed,which can show the evolution trend of social system in a bottom-up way.