| The connections between entities in human society can be described through complex social networks,such as social contact networks,social relationship networks,transportation networks,and so on.The spreading behaviors in these networks have greatly impact on human society,among which the threat of disease spreading,and public opinion spreading are the most prominent behaviors.Outbreaks of SARS and COVID-19 in recent years have huge impacts on the world.With the continuous development of the Internet and social media,the propagation of public opinion on online social networks also has an increasing impact on human society.Current research on social network spreading behavior especially for human contact behaviors,however,is not enough.The factors of disease spreading are difficult to measure accurately,and research about public opinion in different levels with the theories and methods from different fields bring difficulties.It is difficult to support scientific decision-making of public opinion and disease spreading events.To solve these problems,this paper aims to establish an artificial society and carry out large-scale disease spreading and public opinion spreading simulation experiment.In view of the limitations and development trend of the current research about spreading behaviors on social networks,this paper studies the analysis and modeling of social network,the calculation algorithm of large-scale complex network communication,the modeling of data-driven complex network spreading behavior and control measures.The main contents and contributions include the following aspects:(1)A simulation framework for the spreading behavior of complex social networks is proposed.This paper analyzes the characteristics of social network spreading behavior,and the general form of propagation model and social network model,then proposes a simulation framework of spreading behavior on complex social network based on empirical data.The concept of spreading behavior simulation social network is clarified.The concept of network spreading behavior simulation based on empirical data is expounded.The architecture of social network propagation behavior simulation system is proposed,and the structure of the model is explained.(2)Based on the analysis of empirical data,a modeling method of social network with communities of large cities is proposed.Based on the analysis of bicycle travel data and spreading data of online information,the spatiotemporal evolution modeling method of social network is proposed.Finally,combining the temporal and spatial characteristics of the two kinds of social networks,the modeling method of social networks with communities for large cities is proposed.The network can reflect the basic characteristics of social networks,which can be applied to build social networks in a certain region,providing a basis for calculation of diseases or information spreading.(3)A calculation method for spreading behavior simulation on large-scale network based on Spark is proposed.To solve the problem that many social networks have large scale and scale-free characteristics,which will lead to low computing efficiency and difficulty to ensure timeliness,a calculation method of a large scale and complex network spreading based on memory computing is proposed based on Spark which is a big data platform,and the performance is compared with Nepidemix and confirmed.The results show that this method can effectively improve the computing efficiency of spreading behavior simulation computing in large-scale complex networks.(4)Spreading model based on social network and mitigation strategy model based on empirical data are proposed.The spreading models include the rumor spreading model based on the analysis of empirical data of social network,as well as the disease spreading model built based on many related studies about COVID-19.The mitigation strategy models include the gravity immunization strategy model for information spreading and the eight-tuple mitigation strategy model for disease spreading.For spreading model and mitigation model,empirical data analysis is the basis of parameter acquisition.(5)Evaluation of the immunization strategy effect based on rumor spreading model.Based on the small-scale configuration network and real-world information spreading network,the rumor mitigation strategies model are evaluated through simulation experiment,and the advantages of gravity immune strategy are illustrated.At the same time,it is shown that based on multi-agent modeling and complex network modeling theory,information spreading model is built based on empirical data to carry out public opinion mitigation strategy evaluation experiment and provide decision support for public opinion events.(6)Evaluation of the performance of typical pandemic mitigation strategies based on the probability model of disease spreading and the pandemic mitigation strategy model.Based on the artificial social population model,the spreading model and the disease course development model of COVID-19,and the 8-tuple model of pandemic mitigation strategy,the simulation calculation experiment of COVID-19 spreading is carried out in the environment of a typical city,which verifies the rationality of the spreading model and evaluates the performance of typical pandemic mitigation strategies.This article focuses on spreading behavior in social networks.It mainly focuses on the disease spreading and public opinion spreading events with the largest and most destructive capacity.Based on empirical data,the characteristics of social networks are studied,the calculation efficiency of spreading simulation computation are promoted,the disease spreading model of rumors,rumors spreading model and immunization strategy model are enriched and improved,and the disease mitigation strategy model is put forward.The research is of great significance to promote the understanding of spreading behaviors in social networks.It is also helpful for prevention and mitigation of disease outbreak and public opinion events and decision support. |
PDF Full Text Request