Non-parametric Bayesian Estimation of a Spatial Poisson Intensity

From MaRDI portal





The authors introduced a new approach to nonparametric Bayesian estimation of the intensity of a non-homogeneous Poisson process on the real line. The basic idea was to use piecewise constant functions with a random number and random locations of jump times to approximate ``real intensity functions. In this way an intensity defined on a finite interval was parametrized by a finite number of real numbers. The purpose of this work is to develop a similar method for the estimation of the intensity of a spatial Poisson process. The authors use Voronoi tessellations of random point patterns to provide partitions to domains of constant intensity. Hence, step functions are defined through marked point patterns, where the mark of a point gives the intensity value inside the associated Voronoi tile. The authors apply the scheme of conditional autoregression by specifying the conditional distributions of individual intensity levels given the values of all others. The performance of the method is illustrated in examples with both real and simulated data.




Cited in
(37)








This page was built for publication: Non-parametric Bayesian Estimation of a Spatial Poisson Intensity

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4255144)