Research On Cross-network Information Diffusion Model Based On Role Classification

The rise and development of the Internet continues to affect every aspect of human life.With the emergence of many excellent online social platforms,people are active in different social networks,spreading information among the major platforms.For problems such as viral marketing,the evolution and control of public opinion,it will be very limited to consider only a single social network.At the same time,with the advent of the "we media" era,the resource allocation on the Internet is also being adjusted.Unlike the traditional diffusion mode where only official media are available to the public,ordinary netizens can also become the focus on the Internet.Therefore,this paper attempts to explore the identification of key nodes in social networks and the division of users’ propagation roles.On the basis of these,the paper studies the process of information diffusion in multiple social networks,so as to further understand the diffusion mechanism of network information.This is not only of commercial value,but also of great significance to the guidance and control of public opinion.The research of this paper is divided into two parts: the identification of the key nodes in the complex network and the establishment of the information diffusion model across the network.The details are as follows: 1.For the subsequent analysis of user roles,the identification of key nodes was first studied.By improving the classical Kshell decomposition algorithm,the multi-order neighbor property of the node is used to evaluate the influence of the node by combining the global property and local property of the neighbor,so as to sort the user nodes that play an important role in information diffusion.The accuracy and effectiveness of the proposed method were evaluated by calculating the Kendall correlation and imprecision function based on the ranking results obtained from the experiment and the propagation capability obtained from the SIR propagation model.Monotonicity and complementary cumulative distribution function curve is used to measure the degree of differentiation of sorting node.This identification method is proved to be effective in selecting the seed nodes in the influence maximization problem.Experimental results on a variety of real data sets verify that the proposed method is superior to other comparison methods.2.Based on the identification method of key nodes proposed in this paper and other network attributes of nodes,the propagation roles of user nodes in social networks are analyzed,and the characteristic quantities of opinion leaders and bridge characteristics are defined.Users are divided into opinion leaders,bun users and ordinary users.The differences in user roles will exert different degrees of influence on the dissemination of information to others;Crossnetwork entity users are the common users of two networks.This paper analyzes the characteristics of such users and the self-propagation behaviors that increase the flow of information between networks.The similarity of time and consistency of interest generated by cross-network users’ propagation behaviors in the two networks are used to measure the self-propagation probability of users.Based on the analysis of the role of node propagation and the behavior of cross-network propagation,a cross-network information diffusion model is proposed.The paper analyzes experimentally the key role of cross-network users’ selfpropagation in the diffusion of information,and verifies the effectiveness of the proposed model.