A Rumor Propagation Model Based on User Cognition and Evolutionary Game

In social networks, studying rumor propagation patterns is essential for curbing the spread of rumors. Given the coexistence and conflict of multiple-type rumor information, as well as users’ cognitive differences, this article presents a rumor propagation model grounded in user cognition and evolutionary game theory. First, considering the potential impact of social relationships between users on rumor propagation, the KD-Tree algorithm is employed to uncover hidden connections between users, thereby enriching the topology of the user’s social network. Second, a user behavior driving mechanism for rumor, anti-rumor, and motivation-rumor types is constructed based on evolutionary games to reflect the interactive and strategic nature of users’ responses. Moreover, the Lotka-Volterra equation is utilized to explore the dynamic game of multi-type rumor information and the cognitive process of users. Finally, to address differences in users’ cognition, this article introduces the anti-rumor trust state A and the motivation-rumor trust state M , which arise from users’ exposure to multiple types of rumor information. Based on these trust states, a rumor propagation model, SIAMR, is constructed using user cognition and evolutionary game theory. Experiments demonstrate that the model accurately captures the dynamic interactions between multi-type rumor information and the transmission process of rumor topics in social networks. The proposed model integrates cognitive psychology with a strategic interaction framework, offering a more realistic representation of rumor propagation behavior in the real world. Experimental results reveal that SIAMR improves prediction accuracy by 14.23% over baseline models in simulating the dynamics of multiple types of rumors, effectively capturing users’ cognitive influences and the mechanisms of information competition.