ESTIMATING THE MEAN OF INVERSE GAUSSIAN DISTRIB WTION WITH KNOWN COEFFICIENT OF VARIATION UNDER ENTROPY LOSS

Abstract

An estimation problem of the mean µ of an inverse Gaussian distribution
IG(µ, C µ) with known coefficient of variation c is treated as a decision problem
with entropy loss function. A class of Bayes estimators is constructed, and
shown to include MRSE estimator as its closure. Two important members of
this class can easily be computed using continued fractions