Mining Social Networks for Dissemination of Fake News Using Continuous Opinion-Based Hybrid Model

Maneet Singh, S. R. S. Iyengar, Rishemjit Kaur

Abstract: The entire world is confronting the challenge of fake news disseminated online, as its consequences could be exceptionally catastrophic. In this paper, we have proposed a hybrid model that integrates the opinion evolution process with the propagation of fake news. The level of extremity in opinions, the amount of support from social connections and the social influence were used as the major design considerations in modeling the spread of fake news. As polarized opinions on social media often lead to polarized networks, the proposed model was utilized to study the effect of evolving opinion on the spread of fake news on polarized networks of varying degrees. Our findings suggested that there are more users involved in sharing fake news in the presence of a highly polarized network. Moreover, the tendency of a user to adapt the opposing opinion seems to be correlated with the exposure of fake news. Besides this, we also assessed the consequences of the spread of fake news on the user’s opinion and found that the users that are mainly influenced are the ones having an unclear stance towards a given issue. Overall, our proposed model highlights the interrelation between fake news and the opinion evolution on social networks.