English

Modularity of Erd\H{o}s-R\'enyi random graphs

Combinatorics 2022-12-22 v2 Statistical Mechanics Social and Information Networks

Abstract

For a given graph GG, each partition of the vertices has a modularity score, with higher values indicating that the partition better captures community structure in GG. The modularity q(G)q^*(G) of the graph GG is defined to be the maximum over all vertex partitions of the modularity score, and satisfies 0q(G)<10\leq q^*(G) < 1. Modularity is at the heart of the most popular algorithms for community detection. We investigate the behaviour of the modularity of the Erd\H{o}s-R\'enyi random graph Gn,pG_{n,p} with nn vertices and edge-probability pp. Two key findings are that the modularity is 1+o(1)1+o(1) with high probability (whp) for npnp up to 1+o(1)1+o(1) and no further; and when np1np \geq 1 and pp is bounded below 1, it has order (np)1/2(np)^{-1/2} whp, in accord with a conjecture by Reichardt and Bornholdt in 2006. We also show that the modularity of a graph is robust to changes in a few edges, in contrast to the sensitivity of optimal vertex partitions.

Keywords

Cite

@article{arxiv.1808.02243,
  title  = {Modularity of Erd\H{o}s-R\'enyi random graphs},
  author = {Colin McDiarmid and Fiona Skerman},
  journal= {arXiv preprint arXiv:1808.02243},
  year   = {2022}
}

Comments

35 pages, to appear in Random Structures and Algorithms