SoftNet: A Package for the Analysis of Complex Networks

Abstract

Identifying the most important nodes according to specific centrality indices is an important issue in network analysis. Node metrics based on the computation of functions of the adjacency matrix of a network were defined by Estrada and his collaborators in various papers. This paper describes a MATLAB toolbox for computing such centrality indices using efficient numerical algorithms based on the connection between the Lanczos method and Gauss-type quadrature rules.

Keywords

• complex network analysis
• centrality measure
• matrix function
• Lanczos algorithm