Mining triadic closure patterns in social networks

Abstract

A closed triad is a group of three people who are connected with each other. It is the most basic unit for studying group phenomena in social networks. In this paper, we study how closed triads are formed in dynamic networks. More specifically, given three persons, what are the fundamental factors that trigger the formation of triadic closure? There are various factors that may influence the formation of a relationship between persons. Can we design a unified model to predict the formation of triadic closure? Employing a large microblogging network as the source in our study, we formally define the problem and conduct a systematic investigation. The study uncovers how user demographics and network topology influence the process of triadic closure. We also present a probabilistic graphical model to predict whether three persons will form a closed triad in dynamic networks. The experimental results on the microblogging data demonstrate the efficiency of our proposed model for the prediction of triadic closure formation.