CMStatistics 2020: Start Registration
View Submission - CMStatistics
Title: Modeling heterogeneous networks in the presence of covariates Authors:  Swati Chandna - Birkbeck, University of London (United Kingdom) [presenting]
Abstract: Many applications routinely observe covariates at node and dyad level in addition to pairwise interactions between the agents of interest. A nonparametric approach to modeling unlabeled networks is offered by the graphon function. Recently, there has been a growing interest on the problem of graphon estimation as well as its application to important problems such as bootstrapping networks, testing for equivalence of network distribution using subgraph counts, estimation of missing links. Existing histogram approximations to graphon function are not designed to estimate heterogeneity across the full network. We will discuss an approach to modeling heterogeneity in network data using covariates via the graphon model.