Abstract

In this paper we consider the problem of best linear unbiased estimation and best linear invariant estimation of the common scale parameter of a normal and double exponential distributions using some functions of spacings of all observations of individual samples. We have also proved a sufficient condition for the nonnegativity of the common scale estimator obtained by the above method. Further we have obtained necessary and sufficient condition for the derived estimators to be constant multiple of the sum of first and last spacings of pooled sample.

Author: R. S. Priya and P. Yageen Thomas

Received on: August, 2011

Accepted on: December, 2011