Asymptotic distributions of estimators for posterior probabilities in a classification model with both continuous and discrete variables
A.W. Ambergen · Centrum Wiskunde & Informatica (CWI), the national research institute for mathematics and computer science in the Netherlands · 1984
This paper is devoted to the asymptotic distribution of estimators for the posterior probability that an observation vector originates from one of k populations.The estimators are based on training samples.The random vectors contain both continuous and discrete variables.Observation vector and prior probabilities are regarded as given constants.The continuous part of the random vector has conditional on the discrete part a multivariate normal distribution.Several assumptions about homogeneity of the dispersion matrices are considered.