Bayesian estimation for the mixing proportion in classification and discriminant with two populations

Abstract

The article presents a problem in classification with two populations H1 and H2 by Bayesian method, which is building posterior probability distribution function for v, proportion of H1 in H3 (the mixing of H1 and H2) using its truncated (0,1) prior distribution and observations from H3. The cases when v has prior beta, normal and exponential distributions are studied completely.