Statistical significance of communities in networks

Preprint
Publisher Website
Google Scholar
  • 21 July 2009
Abstract
We introduce a new measure aimed to quantify the statistical significance of communities in networks. Communities are seen as groups of nodes which maximize the densities of internal connections. Extreme Statistics is used to predict the statistics associated with communities, including confidence intervals. This allows us to quantify the statistical significance of communities in networks, detecting "artificial" communities arising as structural fluctuations in random graphs without intrinsic structure. The method is successfully applied in the case of real-world networks for the evaluation of the significance of their communities.

This publication has 0 references indexed in Scilit: