Bayesian analysis of spatial point processes in the neighbourhood of Voronoi networks

  • Authors:
  • Øivind Skare;Jesper Møller;Eva B. Vedel Jensen

  • Affiliations:
  • Department of Biostatistics, University of Oslo, Oslo, Norway;Department of Mathematical Sciences, Aalborg University, Aalborg, Denmark;Department of Mathematical Sciences, University of Aarhus, Aarhus, Denmark

  • Venue:
  • Statistics and Computing
  • Year:
  • 2007

Quantified Score

Hi-index 0.00

Visualization

Abstract

A model for an inhomogeneous Poisson process with high intensity near the edges of a Voronoi tessellation in 2D or 3D is proposed. The model is analysed in a Bayesian setting with priors on nuclei of the Voronoi tessellation and other model parameters. An MCMC algorithm is constructed to sample from the posterior, which contains information about the unobserved Voronoi tessellation and the model parameters. A major element of the MCMC algorithm is the reconstruction of the Voronoi tessellation after a proposed local change of the tessellation. A simulation study and examples of applications from biology (animal territories) and material science (alumina grain structure) are presented.