Research On The Spread Of Rumors In Social Networks

In recent years,the rapid development of social networks has greatly changed people's lifestyles.People can express their opinions on social networks,record their lives,and have more diversified channels for obtaining information and knowledge.But at the same time,because the dissemination of this information has not been tested by traditional media,the existence of rumors has become more widespread.The large-scale dissemination of rumors will not only affect people's social network experience,but also seriously threaten social stability,so research how to control rumors in online social networks is an important topic.Many researches on the spread of rumors in social networks ignore the self-control ability of social networks.This thesis discusses how social networks can be implemented through KOL(Key Opinion Leader)nodes that rely on the existence of social networks without intervening in external forces and how to use limited costs to improve KOL's self-control ability.This thesis is mainly divided into two parts,the social network rumor propagation model based on KOL nodes and the cost-constrained rumor control problem based on KOL nodes.The work of this thesis based on two research points is as follows:This thesis believes that social networks have self-control capabilities.When rumors spread in social networks,KOL nodes will judge the authenticity of the rumors based on their knowledge and then influence their fans,so that nodes that refute rumors will appear in the network.Therefore,we propose a rumor spreading model that includes S(Susceptible)node,C(Credulous)node,D(Dubious)node,and R(Removal,)node.And then this thesis names it SCDR model.Among them,the D nodes are generated after the KOL nodes are activated and decided to refute the rumor.This thesis defines two parameters related to the attributes of KOL nodes.One is the activation rate ?,which indicates the probability of whether a KOL node will be activated in the spread of rumors,and the other is the accuracy rate(p,which represents the probability that the information received by the KOL node can be accurately judged as a rumor when it is activated.If it can be judged as a rumor,the KOL node will be converted into a rumor refuting node and influence its fans to refute the rumor.Subsequently,in order to verify the SCDR model,in addition to choosing to simulate in the self-built network,we also used a python crawler to crawl Sina Weibo and obtained a dataset containing 520,000 users,and simulated the spread of rumors on this data set to verify our proposed model.The experimental results prove that social networks have self-control capabilities,and the existence of KOL nodes is of great significance for controlling the spread of rumors.This thesis proposes a rumor control method based on KOL nodes under cost-constrained conditions.This thesis analyzes the traditional node value and node cost calculation methods,and because KOL nodes are important in the spread of social network rumors,this thesis proposes KOL nodes based rumor control with limited cost(KNRCLC)problem.This thesis models the above problems by effectively defining the value calculation and cost calculation of KOL nodes.Then,this thesis proposes three algorithms based on the KNRCLC problem.The three algorithms select KOL nodes from different dimensions as seed nodes to dispel rumors.Experiments show that compared with other algorithms,MCFKN,MIFKN and MICKN algorithms have good performance when the cost is fixed.