Hub-Based Community Finding

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.

4 pages, 3 figures, 1 table

Cited by in corpus (7)

• Community detection in graphs
• Characterization of complex networks: A survey of measurements
• Community landscapes: an integrative approach to determine overlapping network module hierarchy, identify key nodes and predict network dynamics
• A Local Method for Detecting Communities
• Community Detection Using A Neighborhood Strength Driven Label Propagation Algorithm
• Computing communities in large networks using random walks (long version)
• Communities as Well Separated Subgraphs With Cohesive Cores: Identification of Core-Periphery Structures in Link Communities