The Social Influence Analysis Based On The Range Of Second-Order Neighborhood

With the development of Internet and Web2.0,the research of social influence has been widely used in the fields of marketing,product promotion,information monitoring and virus control.Therefore,social influence analysis has important meaning in many fields.Nowadays,the rapid development of various social websites provides a lot of real data for social network research.So,it becomes a problem to solve how to quickly and effectively find important users and influence more people.This article analyses from two aspects of social impact mainly based on the user in the second-order neighborhood.One is the sorting problem of influence.The other is the maximum communication problem of influence,the main research as follows:Firstly,this paper introduces the research status at home and abroad from the issue of influence ranking and maximization,describes the definition and related concepts of social network and complex network and the theoretical basis of social network.Then,this paper briefly summarizes the direction of research and exploration of social network in the future.Secondly,this paper presents a influence sorting algorithm with the node for PSNR,which is based on the second-order neighbor range.It solves two problems which the Degree Centrality algorithm,the PageRank algorithm and the heuristic phase of Hybrid algorithm ignore.One is the end nodes in the network,for zombie users.The other is the situation that users which have second-order fans have a greater influence.So,this paper is more accurate and authenticity for the selecting of important nodes.Then,for the issue of maximizing influence,this paper proposes an improved hybrid algorithm,OHG,which aims at the defect of greedy algorithm and hybrid algorithm.It changes the selection for seed nodes in the heuristic stage.And in the greedy stage,it use a new method.So it is more accurate and more efficient for operation to solve the problem of greater impact.Finally,we experiments with traditional algorithms on real social network data sets,and verifies its accuracy and efficiency through comparative analysis.