A methodology for generating time-varying complex networks with community structure

Abstract

There is a demand of benchmark networks for testing community detection algorithms in dynamic scenarios. For generating these benchmarks is necessary to have a methodology able to create controlled networks that simulate the natural behavior of communities over time. This work aims to fill this gap by presenting a methodology for generating dynamic complex networks with community structure. Computer experiments show that our methodology can, starting from an initial network, evolve its communities over time while the overall modular structure of the network is preserved. © 2014 Springer International Publishing.