Finite Mixtures of ERGMs for Ensembles of Networks
share
Ensembles of networks arise in many scientific fields, but currently there are few statistical models aimed at understanding their generative processes. To fill in this gap, we propose characterizing network ensembles via finite mixtures of exponential family random graph models, employing a Metropolis-within-Gibbs algorithm to conduct Bayesian inference. Simulation studies show that the proposed procedure can recover the true cluster assignments and cluster-specific parameters. We demonstrate the utility of the proposed approach using an ensemble of political co-voting networks among U.S. Senators.
READ FULL TEXT