Linear Estimation Of Standard Deviation Of Logistic Distribution: Theory And Algorithm

Abstract

The paper presents a theoretical method based on order statistics and a FORTRAN program for computing the variance and relative efficiencies of the standard deviation of the logistic population with respect to the Cramer-Rao lower variance bound and the best linear unbiased estimators (BLUE's) when the mean is unknown. A method based on a pair of single spacing and the 'zero-one' weights rather than the optimum weights are used. A comparison of an estimator based on four order statistics with the traditional estimators is considered.