Research On Heterogeneous Social Contagion Dynamics In Complex Networks

In recent years,with the rapid development of the science and technology,the network age has been coming and people are surrounded by kinds of spreading on real world networks.There are a lot of software in the major application platform promoted to be downloaded,a variety of fashion elements prevalent in the streets and many rumors or public opinion widely spreading in major sites and so on.On one hand,the spreading behaviors can improve people' life,on the other hand,they are harmful to human interests.So it is one of the major research topics to study social contagion mechanism and make appropriate strategies for the early warning and control today.At present,the heterogeneity of individuals' adopting behavior,different kinds of network structure,the memory of individuals,the non-redundant transmission and the propagation of non-complete neighbors in the process of social contagion are not yet been synthetically considered.Therefore this paper takes a series of researches on social spreading dynamics and immune based on complex network theory and achieve certain results:(1)We propose the heterogeneous social contagion model and carry out the theoretical analysis.We call it as heterogeneity integrating non-redundant propagation,non-Markov characteristics,non-complete neighbors' propagation,heterogeneous individuals' adopting threshold and different structural characteristics on networks.Then we put forward the heterogeneous social contagion model and improve the theory of edge division,forming the heterogeneous edge division theory to take the steady state analysis and the outbreak threshold analysis.Simultaneously,the theory is fitted for the subsequent experimental simulation.(2)We form a large number of heterogeneous social contagion experiments and theoretical fittings.In order to understand the influence of heterogeneity on social contagion,this paper carries out a large number of experiments to simulate the social contagion process based on the heterogeneous social contagion model in the unrelated ER networks and SF networks and makes the theoretical fitting.It is found that,with the heterogeneity of the adopted threshold stronger,the behavior propagation is suppressed firstly and then promoted when the adoption threshold is completely positive correlated with the degree,and that the behavior is more favorable when completely negative correlated with the degree.Besides,when the adoption threshold is partial positive correlated with the degree,the more irrelevant,the behavior is more favorable,and when partial negative,the more irrelevant,the behavior is more inhibited.The final adoption ratio increases with the network transmission rate and there will be a phase transition,secondary phase transition and mixed phase transition phenomenon.(3)We use degree centrality,approaching centrality and k-core centrality to immune nodes to observe the immune effect on heterogeneous social contagion.It is found that,in the ER network and the SF network,when the adoption threshold and the degree are uncorrelated or completely negative correlated,it is suggested that degree centrality immunity should be selected.When the adoption threshold is completely positive correlated with the degree,if the relationship is weaker,degree centrality immunity is more suitable,but if stronger,it is suitable to select k-core centrality for ER network and the centrality immune is unsuitable for SF network.