Non parametric Bayesian inference for inhomogeneous pairwise interaction point processes

Inhomogeneity in a pairwise interaction point process is modelled by a location dependent first order interaction term, using a Bayesing setting with a shot noise process prior. Simulation of the posterior is done using a new auxiliary variable technique which avoids the unknown normalizing constant in the likelihood. Supported by MaPhySto and the Danish Natural Science Research Council