Research On Rumor Source Detection

The rapid development of the network facilitates our life,but it brings us many security risks.Network rumors are one of the security risks that threaten the stability of the network.As a kind of unauthorized information,network rumors may not only mislead users in the network,but also undermine the stability of the network,and even bring unpredictable risks to the society.Therefore,it is necessary to develop practical means to identify rumor source in order to mangage the risk which comes from rumors.This paper studies the rumor propagation and the detection problem of single rumor source based on the susceptible – exposed – infected-recovered(SEIR)model.The main contributions of this paper are as follows:1.Analyzes the spread of rumors in different network structures based on the SEIR propagation model.Firstly,the dynamic equation of rumor propagation is established based on the SEIR propagation model.Then,the relationship between the density of nodes in different states and the propagation of rumors is analyzed based on the characteristics of homogeneous network and heterogeneous network.It is concluded that rumor propagation is related to the network structure,the average degree of nodes and the probability of receiving rumors for nodes.Finally,the conclusion is validated through simulation experiments,and the immunization analysis is given.2.Investigates the detection problem of single rumor source based on the SEIR propagation model.With an observed snapshot in the network,it is easy to get the relevant graph topology and all infected nodes,while it is difficult to distinguish which nodes in susceptible,exposed or recovered status,the goal of this paper is to identify effectively the rumor source based on the observed snapshot and graph topology.The concept of the optimal infection process is firstly proposed and then an estimator is derived for the rumor source based on this optimal infection process.Subsequently,it is proved that this estimator matches the rumor source with a high probability.Finally,the detection probability of rumor source of our scheme is evaluated based on different types of networks.The results suggest that our proposed scheme outperforms the traditional rumor centrality heuristics.