Supervised classification of the scalar Gaussian random field observation

Abstract

The problem considered is that of classifying the scalar Gaussian random field observation into one of two populations specified by different parametric mean models and common covariance function for given classified training sample. This approach is usually called as supervised classification. The approximation of the expected error rate associated with plug-in Bayes discriminant function (PBDF) are obtained. This is an extension of the previous one to the case of complete parametric uncertainty.