Random walks on random networks of cliques: Inferring the network structure

Abstract

We study the properties of discrete-time random walks on networks formed by randomly interconnected cliques, namely, random networks of cliques. Our purpose is to derive the parameters that define the network structure—specifically, the distribution of clique size and the abundance of inter-clique links—from the observation of selected statistical features along the random walk. To this end, we apply a Bayesian approach based on recording the times spent by the walker inside successively visited cliques. The procedure is illustrated with some numerical examples of diverse complexity, where the relevant structural parameters are successfully recovered.