The Research Of Rumor Propagation Mechanism On Microblog And The Simulations

Microblog has become a platform of mass public opinion, but it also become a factory of producing the rumor. While people believe rumors blindly makes persons take wrong action, it will have a negative impact on the normal society. This paper used some research methods such as mathematical modeling, multi-agent simulation and case analysis to comprehensive study the rumor evolution, the conclusion may provide some effective advice to monitor and manage the microblog.First, based on the analysis of the characteristics of microblogging, the epidemic model was constructed to describe the rumor propagation on micrblog. We expanded the people to the five class (ignorant, infected, contacted, exhausted, resistant), add the coefficient of interest decay to the model. We made comparison to the simulation data and the data from two real cases and found the model can be fit reality better.Second, we extended the online propagation to the offline spread, even considering the action which caused by the trust of the rumor. We build a composite network which included a small-world network of online information spread and a regular network of offline behavior diffusion. According to the SIR model, we considered the interest decay of online information spread and social reinforcement of offline behavior diffusion. Then we studied the impact on the rumor spread process when authoritative information get involved, consider four factors:intervening time, the initial intervention efforts, individual willingness to believe authoritative information, the individual’s willingness to disseminate authoritative. By comparative analysis of the four factors, we found some strategies to control the rumors effectively.Finally, we studied that rumors and its corresponding anti-rumors compete in the communication process for more people’s trust. Based on the Lotka-Volterra in the ecology, we constructed the model of the dissemination of rumors and anti-rumors, solved its equilibrium points and stability, then we found the final stable state just related with the influence coefficients between rumors and anti-rumors. We chosen two real cases on the Sina microblog and took some data, then used Eviews to fit the model.