Detecting Political Secession of Fragmented Communities in Social Networks via Deep Link Entropy Method

Abstract

Breakdown of global connectivity in social networks trough disintegration of fragmented but interacting communities leading to political secession is a major source of forming and strengthening echo chambers and political polarization. Hence, quantifying the significance of each edge (the connection or relationship between two particular nodes, for example two friends on Facebook, or two follower/followed accounts in Instagram or Twitter) from the perspective of global connectivity is a crucial problem in online political communication studies. Among the existing methods for quantifying the edge significance in complex social networks, link entropy (LE) has been a very successful one, which takes into account the two nodes’ (making up that particular edge) uncertainties of belonging to different communities. Considering also the contribution of the uncertainties of the adjacent nodes of those two particular nodes, we recently proposed the deep link entropy (DLE) method. In this work, we examine the political secession of disintegrating communities. In particular, we study complex social networks consisting of multiple communities which are in direct or indirect interaction through bridging individuals. We consider scenarios where those bridges are lost through unfollowing or unfriending an individual belonging to a different community. We show that the DLE method detects the community disintegration with a high performance. We discuss DLE method’s contribution to social network and online political communication studies, in particular examining the online political secession.



Author Information
Fatih Ozaydin, Tokyo International University, Japan
Seval Yurtcicek Ozaydin, Tokyo Institute of Technology, Japan

Paper Information
Conference: MediAsia2021
Stream: Social Media and Communication Technology

This paper is part of the MediAsia2021 Conference Proceedings (View)
Full Paper
View / Download the full paper in a new tab/window


Comments & Feedback

Place a comment using your LinkedIn profile

Comments

Share on activity feed

Powered by WP LinkPress

Share this Research

Posted by amp21