The Research For Minimize The Local Influence Of Negative Information In Social Network

With the rapid development of the Internet,especially the rapid rise of Web2.0,there are a large number of online social networking applications which have different purposes and areas emerge,such as blog,micro-blog and Social Network Service(SNS)websites.SNS websites have developed rapidly in recent years and become an indispensable platform for users to collect and disseminate information,express their views and communicate with others.However,the negative impact of SNS cannot be ignored.Rumors and defamatory information spread willfully,which has a negative effective on society even broken the social order.In order to give full play to the positive role of SNS in information dissemination and promote the healthy development of SNS,this paper has done some relevant theoretical research and work in the negative control information on social networks area,which has some significance for the real online social networks to control the spread of negative information.Based on this,the paper carried out the following work:1.We proposed a method based on blocking the network edge to minimize the impact of negative information.Firstly,we use the directed graph to show the spread of information in the social network and use the greedy algorithm to find the k edges in the directed graph,so that the negative information affected areas is minimized when the k edges are removed,where k is a positive integer;and the k edges cut will make the smallest range of negative information dissemination(infected nodes least).2.We proposed a method based on topic model to minimize the impact of negative information.Specifically,for a social network in which the malicious information has erupted for a while,from the perspective of the topic model,we can calculates the probability distribution of negative information and historical information on each edge probability distributions by the topic model;we calculated the distance between the probability distribution of the negative information and the probability distribution of the historical information on each edge,which is also called the KL divergence;calculate the corresponding calculation parameters,then place them in descending order and remove the top k nodes,so the negative information propagation range is greatly reduced.3.The big Data era has arrived,and the scale of the social network become increasingly larger,the two control methods which are proposed previously are no longer applicable for large scale networks,so we developed a new method of graph partitioning algorithm(MSP),dividing large networks into a small diagram of the networks,and then do the algorithms in each small network to control the impact of the negative information to minimization.