On Maximum Likelihood Estimation of a Spatial Poisson Cluster Process

John M. Castelloe and Dale L. Zimmerman

Technical Report #312, February 2002

Department of Statistics and Actuarial Science, University of Iowa

Abstract:  The likelihood function of a bivariate Poisson cluster process is derived in closed form. Although maximum likelihood estimation in this setting is quite computationally intensive, an example demonstrates the successful convergence of a maximum likelihood estimation procedure for a small data set.

Keywords:  Bivariate normal distribution, Likelihood, Maximum likelihood estimation, Nelder-Mead simplex, Poission cluster process

Full manuscriptPDF, zipped postscript