A Correlative Study of Centrality Measures across Real-World Networks

Abstract

Centrality measures have evolved over the years and used over a variety of networks. Being one of the most basic measures to identify important nodes in ever-increasing modern-day networks, it has been manipulated and modified in every way possible to fit the requirement of the network and the way important nodes are perceived in it. Different centrality measures which seem to perform quite differently on a theoretical basis, provide similar results when applied to real-life networks. The centrality measures are studied based on the approach used, application areas, performance and measure the correlation among 14 centrality measures across 12 network topologies using Pearson, Spearman and Kendall correlation coefficients.