Discovering communities by information diffusion

Discovering underlying communities in networks is an important task in network analysis. In the last decade, a large variety of algorithms have been proposed. However, most of them require global information or a centralized control. Those algorithms are infeasible in large-scale real networks due to computation and accessibility. In this paper, we propose a novel decentralized community detection algorithm based on information diffusion. We believe information diffusion in human society can allow us to understand the emergence of community structure. Being able to find out some critical nodes which play an important role in the formation of a community is an important byproduct for our algorithm. Experiments on various networks, including benchmark networks and synthetic networks, show that it is comparable to three decentralized algorithms and two representative centralized algorithms, in terms of stability and accuracy.