A crucial users dynamic discovery model based on rumor and anti-rumor

Rumor propagation in social networks can be challenging to predict accurately, and pinpointing crucial users is essential for analyzing rumor spread. Existing research exhibits shortcomings in three areas. The first is modeling the coupling of multiple trust relationships and dynamic influence; the second is quantifying the impact of rumor-anti-rumor dynamic interactions; the third is tracking dynamically changing influence in domain-specific propagation. To address these challenges, a crucial users dynamic discovery model based on rumor and anti-rumor is proposed. Firstly, to address the complexity of rumor propagation network topology, explicit/implicit relationship networks are analyzed by integrating topological connectivity and historical interaction data, and a multidimensional trust-influence matrix is constructed. Secondly, focusing on the dynamic game nature of rumors vs. anti-rumor messages, an evolutionary game theory framework is introduced. Competitive dynamics between rumor and anti-rumor propagation are quantified, user behavior patterns are integrated with strategy adaptation, and crucial users’ dynamic evolution is captured. Finally, to address dynamic shifts in crucial users’ domain influence, Latent Dirichlet Allocation is used to extract thematic patterns from rumor data, and domain representation is enhanced. Graph convolutional networks are also employed to combine dynamic influence metrics with domain characteristics for crucial user analysis. Experiments show the proposed model effectively identifies crucial users in rumor and anti-rumor dissemination, reflects multi-layered user trust relationships and their dynamic game dynamics, and helps authorities debunk rumors precisely, boosting social media and public opinion management efficiency.