A1441
Title: Proxy methods for factor copula-based clustered data with random covariates
Authors: Bruno Remillard - HEC Montreal (Canada) [presenting]
Pavel Krupskiy - University of Melbourne (Australia)
Bouchra Nasri - Universite de Montreal (Canada)
Abstract: A proxy method is proposed for the estimation of parameters and latent factors for clustered data based on factor copulas, under the assumption that the number of observations per cluster is very large. The model can also incorporate random covariates, and the conditional distribution of the response variable, given covariates, is a parametric family of continuous or discrete distributions, while the effect of a common latent variable pertaining to a cluster is modelled with a factor copula. As a by-product, a goodness-of-fit test is also proposed. This extends previous work on similar models where the number of observations per cluster was assumed to be small.
