Ternary interaction evolutionary game of rumor and anti-rumor propagation under government reward and punishment mechanism

To aid in designing effective rumor control strategies, combined with system dynamics, the paper proposed a ternary interaction evolutionary game model of participants of rumor and anti-rumor propagation in social network platform, involving the media, government, and netizens. In this model, rules are established to classify source microblogs, comments, and forwards on the real social network platform Sina Weibo as rumors or anti-rumors. It quantifies the ambiguity and severity of rumors, the certainty and evidentiality of anti-rumors, the sentiment of society, and the influence of users, embedding them into the evolutionary game model. Sensitivity analyses were carried out based on the real data from social network platform Sina Weibo to investigate how the exogenous variables affect different evolutionary stable strategies. Experiments demonstrate that the government could adopt loose regulation coupled with a strict but reasonable reward and punishment mechanism, which can promote the media to make true reports and relieve its pressure of rumor regulation. And the cost of the government’s strict regulation and netizens’ anti-rumor propagation have significant impact on their choices when the government chooses loose regulation as its evolutionary stable strategy, while they have a tiny impact on their choice when the government chooses strict regulation as its evolutionary stable strategy. But reducing the cost of media and netizens to search for dispelling evidence can encourage them to propagate anti-rumor information in different situations.