Quick inference for complex Cox process models with different types of inhomogeneity and aggregation

Jiří Dvořák

Inhomogeneity and aggregation are often observed in spatial point patterns. Such efects may be caused by environmental 
inhomogeneity and/or interactions between points. We discuss why log-Gaussian Cox processes are an appealing class 
of models for statistical inference in this context. We also present fast, moment-based methods for parameter estimation 
(two-step composite likelihood), model reduction (a variant of likelihood ratio test) and model validation (simulation-based 
global envelope testing using suitable functional characteristics).