Hub-Based Community Finding

TL;DR

This paper introduces a hub-based method for community detection in complex networks, utilizing wavefronts from high-degree nodes to identify communities and their boundaries across various network types.

Contribution

It presents a simple, effective approach for community finding that works on different network types and includes boundary detection, demonstrated on multiple datasets.

Findings

Effective in identifying communities in various networks

Applicable to weighted and unweighted networks

Capable of boundary detection between communities

Abstract

This article presents a hub-based approach to community finding in complex networks. After identifying the network nodes with highest degree (the so-called hubs), the network is flooded with wavefronts of labels emanating from the hubs, accounting for the identification of the involved communities. The simplicity and potential of this method, which is presented for direct/undirected and weighted/unweighted networks, is illustrated with respect to the Zachary karate club data, image segmentation, and concept association. Attention is also given to the identification of the boundaries between communities.