More on the unbiased ridge regression estimation

Abstract

Listen

Translate article

This paper considers the performance of the unbiased ridge estimator (Crouse et al. Commun Stat Theory Methods 24:2341–2354, 1995) over the ordinary least squares estimator of the regression parameter using the Pitman’s closeness criterion. Then, we introduce a new variance components estimator based on the unbiased ridge estimator. Furthermore we show that the new variance components estimator has smaller mean squared error than the variance components estimator based on the ordinary least squares estimator. A simulation study has been presented to compare the performance of the estimators, and a numerical example has been proposed to explain the performance of the estimators.