Effect of jackknifing on various ridge type estimators

Abstract

The paper considers various adhoc assumptions to estimate the ridge parameter k, when the mean squared error (MSE) of ridge estimator is less than the ordinary least squares (OLS) estimator. We consider the optimal ridge parameter k in the jackknife technique applied in the ridge regression, which has not been done before. Two stages are involved in the present study. First is a comparison of the two estimators of the jackknifed ridge regression (JRR) and generalized ridge regression (GRR) with the OLS, in terms of the mean squared error. The ridge parameter k used is nonstochastic in the JRR and GRR estimators. Second is an empirical study of the jackknifed ordinary ridge regression (JORR) properties. Present study concludes with a numerical illustration.