A Fast Method for Detecting Communities from Tripartite Networks

Abstract

This paper proposes a fast method for detecting communities from tripartite networks. Our method is based on an optimization of tripartite modularity, and the method combines both edge clustering and Blondel's Fast Unfolding. Experimental results on synthetic tripartite networks show that accurate communities are detected with our method. Furthermore, an experiment on a real tripartite network shows that our method is scalable to tripartite networks of tens of thousands of vertices. To the best of our knowledge, this is the first attempt for analyzing real tripartite networks composed of tens of thousands of vertices.