Influence Nodes Mining Algorithm And Information Dissemination Model In Social Networks

In recent years,online social networks Continuous integration with the traditional social network,with the rapid development of Internet,Renren,Sina Weibo based on the relationship between students or concern that is the representative of this type of integrated social networks,the research of social networks information dissemination also gradually rise.Where the mining of influential nodes in the networks may inhibit or accelerate the dissemination and diffusion of information,and this is the influence maximization problem from another point of view.In social network analysis,the research of the model of information dissemination focus on the forms of communication and diffusion mechanism on information under some kind of social relations.Therefore,the study of social network information dissemination has very important significance to understand the information diffusion problem in reality life.In this paper,the innovation proposed according to the following two aspects of social network information communication problems.Firstly,study the existing three classical algorithms which in social networks node influential mining algorithms.Where the Degree Centrality Algorithm is simple,computational efficiency is relatively high,but the influential correlation of the same node is low due to the single network information that the algorithm considers.While Centrality Algorithm and Closeness Centrality Algorithm is based on the global information that is the shortest path,resulting in the high time complexity,and not applicable to online social networks with a large number of users and relationships.Therefore,this paper proposes a new centrality algorithm:Local-import Centricity Algorithm,by studying the above-mentioned several central algorithm.The algorithm associated with high influential correlation of the same node and relatively low time complexity of the algorithm-Local-import Centricity Algorithm.This algorithm evaluate influence of nodes based on local importance of the node and its neighbor nodes,then simulating the top-k node transmission of infectious diseases on the SIR epidemic mode,verified the proposed centricity algorithm has certain validity and practicability.Secondly,through studying the classical independent cascade model in this paper,obtaining that there are some problems that the model results are not realistic in the dissemination of information diffusion process,and presents a cascade model(TF-IC)based on a time constraints.In this model: First of all,after the impact of a node failed,within a certain time limit will continue to activate the node until the node is activated or a time limit arrived.Then,the model will adjust impact probability according to the result of the spread of positive feedback or negative feedback,regardless of each node is activated successfully or unsuccessfully.For the model of two improvement enables the information to spread more widely.In this paper,proposed model has a good spread through the data verification.The data comparison experimental shows the proposed model has good communication.