Subgraph centrality and walk-regularity

https://doi.org/10.1016/j.laa.2019.02.005Get rights and content
Under an Elsevier user license
open archive

Abstract

Matrix-based centrality measures have enjoyed significant popularity in network analysis, in no small part due to our ability to rigorously analyze their behavior as parameters vary. Recent work has considered the relationship between subgraph centrality, which is defined using the matrix exponential f(x)=exp(x), and the walk structure of a network. In a walk-regular graph, the number of closed walks of each length must be the same for all nodes, implying uniform f-subgraph centralities for any f (or maximum f-walk entropy). We consider when non-walk-regular graphs can achieve maximum entropy, calling such graphs entropic. For parameterized measures, we are also interested in which values of the parameter witness this uniformity. To date, only one entropic graph has been identified, with only two witnessing parameter values, raising the question of how many such graphs and parameters exist. We resolve these questions by constructing infinite families of entropic graphs, as well as a family of witnessing parameters with a limit point at zero.

MSC

05C50
05C75
15A16

Keywords

Centrality
Graph entropy
Walk-regularity
Functions of matrices
Network analysis

Cited by (0)

The work of these authors was supported in part by the Gordon & Betty Moore Foundation's Data-Driven Discovery Initiative through Grant GBMF4560 to Blair D. Sullivan.