Abstract
Echo chambers on social media, considered to be a major factor for accelerating political fragmentation, is becoming apparent as a social problem. An echo chamber environment can be a breeding ground for fake news and hate speech, and may have a negative influence on democratic decision-making. In this paper, we investigate echo chambers by analyzing social media data and build a mathematical model that reproduces their characteristics. The simulation results show that the speed of echo chamber formation is affected by the magnitude of social influence and the frequency of social disconnection or unfollow. Furthermore, the social data analysis revealed that following a small number of users with opposite political beliefs may encourage information diversity, while mitigating the formation of echo chambers.
