Persistent Roles in Online Social Networks

Matt Revelle, Carlotta Domeniconi, Aditya Johri

Abstract:

Users in online social networks often have very different
structural positions which may be attributed to a latent factor: roles. In
this paper, we analyze dynamic networks from two datasets (Facebook
and Scratch) to find roles which define users’ structural positions. Each
dynamic network is partitioned into snapshots and we independently find
roles for each network snapshot. We present our role discovery methodology and investigate how roles differ between snapshots and datasets.
Six persistent roles are found and we investigate user role membership,
transitions between roles, and interaction preferences.