Reversible Jump MCMC Analysis of Spatial Poisson Cluster Processes.pdfVIP

Reversible Jump MCMC Analysis of Spatial Poisson Cluster Processes JOHN M. CASTELLOE SAS Institute Inc., USA Abstract: A reversible jump Markov chain Monte Carlo (RJMCMC) technique for bivariate normal mixtures with common covariance matrix is developed and applied to a special type of spatial point process, namely the Poisson cluster pro- cess with bivariate normal offspring dispersal. Inference for this type of process focuses on the common covariance matrix of the offspring dispersal distribution. RJMCMC extends the traditional MCMC capabilities by providing for transi- tions between different parameter spaces, which are needed in our situation due to the unknown number of clusters. A new convergence assessment method, applicable to 
 RJMCMC situation in which distinct “models” (in our case, different numbers of clusters) can be identified, is designed and theoretically justified. Output analysis methods are also developed, including anisotropy test- ing/estimation and inference for number of clusters. The RJMCMC technique is flexible and has potential to apply to more complicated spatial point processes, and also other mixture-related problems. Keywords: CONVERGENCE ASSESSMENT, MIXTURE MODEL, SPATIAL POINT PATTERNS, MARKOV CHAIN MONTE CARLO. 1. INTRODUCTION AND PRELIMINARIES In the estimation of properties of mixture components, proper adjustment for uncertainty about the number of components is a difficult endeavor, with adequate solutions appearing only recently. We consider a problem in the analysis of spatial cluster processes that de- mands such a solution and develop a new approach utilizing reversible jump Markov chain Monte Carlo (RJMCMC) methodology. Details not given here can be found in Cas