Inferring Individual Influence in Social Network

Abstract

We study the integration of individuals attributes to infer their influence ability in social network in this paper. The influence between individuals is usually asymmetric and can propagate via edges gradually. We suggest an Influence Factor Graph(IFG) which can integrate different node and edge features into a uniform inferring model. And for each node the model can compute personalized influence ability value. Experiment results in Zarchary and Wikipedia co-editing social networks show that, the model can depict influence reasonably and reveal some interesting social influence rules.