Toward an emotional opinion formation model through agent-based modeling

Abstract

In the last few decades, many opinion formation models have been proposed to describe how opinion interactions among individuals result in different distributions of opinions within social systems. Emotion plays a key role when people try to influence others' opinions, but applying emotion to opinion formation models has attracted little attention. In this paper, we discuss how emotion can affect opinion formation in social systems. We have used the agent-based modeling and simulation approach due to the complexity of the system. For emotion modeling, we have used the circumplex model of affect, a dimensional model comprised of two dimensions: valence and arousal. The idea has been applied to the Deffuant basic opinion formation model, which is a continuous opinion, nonlinear, and discrete time model. The simulation results of the model show how output parameters of the same opinion formation model such as convergence time, opinion distribution, the number of resulting clusters, and the trend of opinions approaching the final distribution of opinions are affected by the emotional behavior of individuals. Therefore, the results lead us to conclude that considering emotional behavior alongside opinion interaction rules could produce more realistic opinion formation models.