Belief model of complex contagions on random networks

We proposed a belief model on random networks to explore the process of opinion dissemination, focusing on 3 significant factors: the inherent beliefs of individuals, the heterogeneity of individual persuasion ability and the dilution effect of neighbor size on neighbors' persuasion. By mean-field approximation approach, the theoretical final fraction of active agents is determined, which agrees well with simulation results in most situations but diverges around critical condition. This divergence is demonstrated to be caused by the heterogeneity of properties of the initial active node and inevitable for any mean solution. As an alternative to predict and control the contagion, we proposed two strategies for selecting the initial active node based on its properties. Their efficiencies are discussed on different networks by simulations. (C) 2020 Elsevier B.V. All rights reserved.

Automated summary

A belief model on random networks was proposed to study the process of opinion dissemination, focusing on individual inherent beliefs, persuasion ability heterogeneity, and the dilution effect of neighbor size. Theoretical final fraction of active agents was determined through mean-field approximation, showing divergence around critical conditions due to heterogeneity of initial active node properties. Two strategies for selecting the initial active node based on its properties were proposed as alternatives for predicting and controlling contagion, with their efficiencies discussed through simulations on different networks.