B1343
Title: Model selection for network data based on spectral information
Authors: Jonathan Stewart - Florida State University (United States)
Jairo Pena - Florida State University (United States) [presenting]
Abstract: A methodology is presented for model selection in the context of modeling network data. Network data, often represented as a graph, consists of a set of pairwise observations between elements of a population of interests. The statistical network analysis literature has developed many different classes of network data models, with notable model classes including stochastic block models, latent node position models, and exponential families of random graph models. We develop a novel methodology that exploits the information in the Laplacian matrix's spectrum to provide a measure of goodness-of-fit of a defined set of network data models to the observed network. We explore the performance of our proposed methodology to popular classes of network data models through simulation studies, and demonstrate the utility in practice by applying our methodology to a collaboration network of network science researchers and the Sampson monk network.
