Title: Estimation of a model for spatial binary data
Authors: Gabrielle Kelly - University College Dublin (Ireland) [presenting]
Abstract: A model for a realization of a binary random field is considered where the correlations satisfy the Frechet-Hoeffding bounds. The binary variables are related to latent variables that have a Matern spatial correlation. Thus, both the marginal means of the binary variables and their spatial distances contribute to their spatial correlations. The model is fitted to TB infection data in cattle herds and wildlife badgers with a fixed censoring time of 2 years. Estimates of the practical range are obtained, i.e. distances at which spatial correlation can be regarded as negligible.