AMLE3D: A computer program for the inference of spatial covariance parameters by approximate maximum likelihood estimation

Abstract

This paper describes AMLE3D, an ANSI Fortran-77 computer program for the estimation of the covariance parameters of a regionalized variable by maximum likelihood estimation (MLE). The main computational drawback of MLE (the method becomes prohibitive when the number of experimental data is large) has been overcome by using the approximate likelihood instead of the complete likelihood. The program assesses the uncertainty of the estimates by assessing the standard error of the sampling distribution. The performance of the program is demonstrated with two data sets. © 1997 Elsevier Science Ltd.