Shrinkage testimators and their applications for the reciprocal of shape parameter of Pareto distribution

In this paper, the estimation for the reciprocal of the shape parameter of the
Pareto distribution (1/a) has been considered. If (1/a_0) be the prior guess value
of (1/a) then shrunken estimator for (1/a) has been proposed and its properties
have been studied. For socio-economic data n is usually large, the normal approximation
to test statistics can be applied (see Rahman et al. (1997)) which gives
equal tail areas on either side of uniformly most powerful unbiased test. By taking
normal approximations to test statistics a shrunken testimator and preliminary
testimator for (1/a) have been proposed and its properties are discussed.

Fulltext