An approach to model log-Gaussian Cox processes with fully non-separable structures
Log-Gaussian Cox processes define a flexible class of spatio-temporal models which allow the description of a wide variety of dependence effects in point patterns where the clustering structure observed can be described by the inclusion of random heterogeneities in an unobservable intensity function. In this work, the analysis of a spatio-temporal point pattern, corresponding to a set of observed forest fires in Nepal, is performed under different degrees of interaction of the spatial and temporal dimensions. The predictive performance is compared graphically using risk maps, whilst global and local weighted second-order statistics are computed to measure the goodness of fit of each model involved in the analysis.
Palabras clave: log-Gaussian Cox process non-separability risk maps spatio-temporal point process spatio-temporal pair correlation function