Uncovering community structure in networks via hybrid clustering using cascading failure dynamics and topological metric functions

Abstract

Many networks have community structure — groups of nodes within which connections are dense but between which they are sparser. While there exists a range of algorithms for community detection in networks, most of them try to discover this important mesoscale structure from a topological point of view solely. Here we develop a hybrid clustering approach for uncovering the community structure in a network using a combination of information on local topology of the network and on the dynamics of the cascading failures. The originality of the proposed approach is that we introduce a novel fusion of the dynamic behaviors of the cascading failures and topological metric functions in the [Formula: see text]th-nearest neighbor density scheme, which integrates both the global and local structural information of a given network for community detection. The experimental results on both artificial random and real-world benchmark networks indicate the effectiveness and reliability of our approach.