Research On Competitive Propagation Of Rumor And Anti-Rumor In Complex Networks

With the development and popularization of the Internet,rumors in the network are becoming more and more widespread.The research on rumor propagation has become a hot spot in the field of complex networks in recent years.Rumor usually does not exist independently,and generally competes with anti-rumor in the network.In order to better describe the competitive propagation of the two in the real world,this thesis studies the modeling of competitive propagation in complex networks and the popularity prediction,mainly including two tasks:First,a rumor and anti-rumor competitive propagation model is built.Relevant studies believe that there is a small group of people who can see through rumors.Therefore,this thesis proposes a UDK-TNB double-layer coupling network model based on the scenario that "the unknown person has a certain probability of turning into a rumor breaker after being exposed to the rumor and begins to spread the rumor-refuting information",including the rumor layer and the anti-rumor layer.This thesis analyzes the dynamic process within and between layers,uses MMCA(Microscopic Markov Chain Approach)to derive the node state transition Markov equations,and finds out the formula for calculating the outbreak threshold of rumor.The effectiveness of the model is proved by comparing the results of MMCA simulation and Monte Carlo simulation.The influence of each parameter of the model on the popularity of information is analyzed through numerical simulation results.The results show that the increase in the generation rate of rumor breaker and the rate of propagation of rumor can effectively restrain the spread of rumor,and the rate of loss of interest in rumors is proportional to the popularity of rumor information.The existence of the rumor breaker mechanism limits the spread of rumors.In order to be closer to the real world,this thesis builds UDK-TNB double-layer coupled AD(Activity Driven)temporal network model,and explores the influence of temporal network characteristics on information popularity.The experimental results show that the lower node contact level and the lower activity heterogeneity of the rumor layer can well hinder the spread of rumors.Second,ASIR(Alpha-SIR)competitive propagation model oriented by rumors’ popularity prediction is built.The model considers the unilateral suppression and mutual suppression between rumor and anti-rumor,and the propagation rate of one information will change dynamically according to the popularity of its competing information,rather than a fixed value.This thesis also extends the ASIR model to a stochastic form.Through numerical simulation experiments,when there is mutual inhibition between rumors and anti-rumor,if the competition intensity of anti-rumor is low,or the anti-rumor appears too late,it may not be able to effectively suppress the spread of rumors.Based on these conclusions,this thesis put forward the key points to improve the effectiveness of immunization strategies,from the three perspectives of the competition intensity,transmission rate and time delay of dispelling rumors.This thesis studies the ASIR model rumor information popularity prediction algorithm,which performs well on both simulated data and real data.Although the ASIR model models the occurrence of information propagation events without considering the specific network topology,the prediction result in the scale-free network still has good accuracy.Finally,this thesis drives the expression of the basic reproduction number of the ASIR model,and proposes a rumor-dispelling mechanism from the perspective of controlling the basic reproduction number.