A framework for evaluating complex networks measurements

TL;DR

This paper introduces a framework for assessing the quality of measurements used in characterizing complex networks, focusing on resolution, degeneracy, and discriminability, and demonstrates its effectiveness with real-world and model networks.

Contribution

It proposes a novel evaluation framework for complex network measurements, emphasizing their effectiveness and discriminability, and compares different measurement types.

Findings

Symmetry measurements outperform concentric measurements in network characterization

The framework effectively evaluates measurement quality in complex networks

Real-world networks are better characterized using symmetry-based metrics

Abstract

A good deal of current research in complex networks involves the characterization and/or classification of the topological properties of given structures, which has motivated several respective measurements. This letter proposes a framework for evaluating the quality of complex network measurements in terms of their effective resolution, degree of degeneracy and discriminability. The potential of the suggested approach is illustrated with respect to comparing the characterization of several model and real-world networks by using concentric and symmetry measurements. The results indicate a markedly superior performance for the latter type of mapping.