A CLASS OF MODIFIED LINEAR REGRESSION ESTIMATORS FOR ESTIMATION OF FINITE POPULATION MEAN

In recent times, a large number of modified ratio estimators are introduced by assumingvarious population parameters are known. In the same direction we have suggested a class ofmodified linear regression estimators which are unbiased. We have derived their variancestogether with the values for which the proposed class of estimators perform better than the usuallinear regression estimator and existing modified ratio type estimators. Further we have shownthat the estimators from SRSWOR sample and the linear regression estimator are the particularcases of the proposed estimators. The performances of these proposed estimators are alsoassessed with that of linear regression estimator and some of the existing ratio type estimators forcertain natural populations available in the literature. It is observed from the numericalcomparisons that the proposed estimators perform better than the existing estimators and linearregression estimator.