Analysis of opinion spreading in signed social networks under the impact of structural balance

Abstract

Social networks have attracted remarkable attention and it is of great importance to understand the process of opinion spreading in popular social networks. However, most research on diffusion cannot be applied directly to investigate social networks, where relationships are heterogeneous and structural balance is a common phenomenon. In this paper, we propose models to characterize the process of opinion spreading in signed social networks under the impact of structural balance. We classify users into different types according to the numbers of their positive links, and define the term user influence to represent the average number of times that users are influenced, which is incurred by a user spreading an opinion. We then propose an approach to analyze the user influence theoretically and the analysis accuracy is verified by simulations. We observe that the user influence increases with user type and also increases with the fraction of negative links in the network if this fraction value exceeds some point. That's to say, negative relationships may enhance opinion spreading if we consider the impact of structural balance, which is an interesting result.