Evaluation And Research On The Impact Of Source Identification Based On Observation Deployment Strategy

In real world,many phenomena can be modeled as the propagation of sources on complex networks,such as the infection process of diseases on interpersonal contact networks and the spread of rumors on social networks.These networks often have the feature of large-scale and complex structures.How to locate the propagation source in a complex network is a challenging task.This task often infers the position of the propagation source based on some observers.Therefore,it is very important to effectively select the observers for positioning the propagation source accurately.This article focuses on the deployment strategy of observers:Firstly,the effect of deployment strategy of observers on the accuracy of source identification is considered.This work analyzed the influence of deployment strategies including random,degree,clustering,eigenvector,closeness and betweenness for estimating the spreading source.In the experiment,it simulated in three kinds of synthetic networks and four real networks using SI propagation model and reverse greedy algorithm.The results show that using eigenvector deployment strategy is more advantageous to improve the accuracy of estimating the spreading source.Secondly,the effect of the deployment strategy of observers based on community structure on the accuracy of source identification is studied.Pinto has put forward a source identification method based on the part of the observer.However,this method is not analyzed the influence of the selection of observers on algorithm accuracy.At the same time,it does not take the influence of the community structure on location the source of diffusion into account.According to these,use the LFR benchmark network and GN networks as the data set,analyzed the influence of the community structure and the motion of the selection of observers on location the source of diffusion.Through a large number of experiments,we found the phenomena: if observers and the source exist in the same community,the accuracy of identifying the propagation source is higher that observers and the source are not in the same community.If observers were randomly selected nodes,as some observers and the source in the same community,others and the source is not in the same community,the possibility of location the source of diffusion is between the two.Therefore,the deployment strategy of observers based on the community structure will affect the accuracy of the propagation source.The choice of deployment strategy of observers is very important to locate of the dissemination source.Firstly,this article deploys observers from the perspective of the nodes' index information in the network,and then deploys observers from another perspective on the network structure.Finally,a large number of experiments verify the importance of deployment strategy of observers.