A MULTI-AGENT SIMULATION APPROACH TO RUMOR SPREAD IN VIRTUAL COMMNUNITY BASED ON SOCIAL NETWORK

In general, rumor spread has certain impact on public safety and it has become a universal phenomenon in virtual community. Based on the data collected from Kaixin website, which is a famous social network service and website launched in 2008 in China, it is briefly analyzed the network properties, such as maximum degree, average degree and so on. Then a multi-agent rumor spread model in virtual community is established based on the analysis of rumor spread pattern. According to the proposed model, some simulations are conducted to show the impact of network structure, tolerance frequency, and believing rate to rumor spread. It is also concluded that in this kind of virtual community, rumor cannot be eliminated and its spread is also insufficient; each node's tolerance frequency can greatly reduce the prevalence of rumor duration; believing rate of each node can change nodes' trusting states to rumor obviously in the spread process. Finally, it is also suggested that some strategies on Internet culture should be adopted to resist on rumor spread in virtual community.