PREDICTIVE ESTIMATION OF FINITE POPULATION MEAN USING GENERALISED FAMILY OF ESTIMATORS

Abstract

Using predictive estimation procedure, an attempt has been made to develop some generalisedfamily of estimators for the finite population mean ¯J of survey variable J in the presence of known auxiliaryvariable K. The proposed class consists of mainly two different types of estimators namely, Ratio andExponential ratio-product type estimator. Theoretical conditions under which the proposed classes areless biased and more efficient then usual unbiased, ratio estimator and estimator due to [2], [12] and [13]have been obtained . It is also evaluated that at optimum values of unknown scalars the mean square error(MSE) of suggested classes tends to the MSE of regression estimator. Finally, these theoretical findings areillustrated by a numerical example