Research About Information Propagation In Social Network

In recent years,with the rapid development of Internet technology,online social networking is gradually changing people's access to information and social way.Facebook,Twitter,and Youtube in the United States,as well as the domestic Sina microblog,QQ,and WeChat,which is the representative of a large number of online social network,have accumulated hundreds of millions of users.Social network with its real-time,functional,social and many other advantages has became one of the most important applications of web2.0 architecture,which has important influences on the current social news events,false rumors,public opinion and other information.The use of social network can create great value for mankind,but at the same time,the advantages of social network may also improper use and cause great harm.Therefore,the research of information propagation is of great significance,and it is a hot field of social computing in recent years.Firstly,this paper introduces the development of social network and its research significance,and lists the basic knowledge of complex network and information propagation.Secondly,according to the characteristics of propagation in social network,based on the traditional Susceptible-Infectious-Recovered(SIR)model,the Susceptible-Disguising-Infectious-Recovered(SDIR)model is established by adding a kind of new node named disguising node.Finally,considering the mutual influence of neighbor nodes,three propagation probability functions are defined to improved SDIR model.The results show,by simulating propagation under different conditions,information can't cover the whole network,and twitter perform better than Sina Micro-blog in efficiency of propagating.Also the initial infection probability have a significant influence in the information propagation.In process of simulating the SDIR model,it is found that the basic propagation probability must be set with the empirical value before propagation for using various propagation models.However,through the analysis,this paper found propagation probability of man-made has great influences on results.Therefore,based on idea of path ranking algorithm in Knowledge Graph,this paper presents a method to calculate influence of nodes by random walk,and the probability of propagation is obtained by normalizing on the basis of influence.Experiments compared different results caused by fixed propagation probability and the probability considering influence of source,and showed calculated probability is more satisfied with truth by validating the algorithm of node's influence.Then calculated probability and SDIR model are combined to further improve the information propagation model in this paper.