Unbiased estimation of distribution densities of sufficient statistics of the inverse Gaussian distribution (Q1269933)

The two-parameter inverse Gaussian distribution is widely used in the construction of probability models of many real processes. Therefore, the construction of unbiased estimators for functions of the parameters and for the density of this distribution is important. The problems of unbiased estimation of some functions of unknown parameters of the inverse Gaussian distribution were considered in several papers. In the present paper, unbiased estimators for the densities of sufficient statistics and likelihood functions are found for the case of inverse Gaussian distributions. We also consider the problem of construction of asymptotically optimal decision rules of group classification.