ABSTRACT
Spatial point process models provide a large variety of complex patterns to model particular cluster situations. Usually, the main tools to compare theoretical results with observations in many varied scientific fields such as engineering or cosmology are statistical, so the new theories and observations also initiated an active use of spatial statistics in these fields. However, due to model complexity, spatial statistics often rely on MCMC computational methods, a backbone for using these techniques in practice.
In this paper we introduce point field models, such as the continuum random-cluster process, the area-interact ion process and interacting neighbour point processes, that are able to produce clustered patterns of great variety. In particular, we focus on computational aspects for simulation and statistical inference while analyzing the flexibility of these models when used in practical modelling.
