Empirical Bayesian classification for observations with admixture
Olena Sugakova · Theory of Probability and Mathematical Statistics · 2012
We consider the problem of classification of an object by using observations after a numerical characteristic under the assumption that each object belongs to one of the two given classes. The distribution of the characteristic is unknown for objects of the first class but is assumed to be symmetric. The distribution for the second class is known. We construct an empirical Bayesian classifier and prove a result concerning the asymptotic behavior of the error probability.