Generalized ridge and principal correlation estimator of the regression coefficient in growth curve model

Abstract

This paper develops a new biased estimator—generalized ridge and principal correlation (GRPC) estimator of the regression coefficient in the growth curve model. The properties of the proposed estimator are shown to be superior to those of least squares (LS) estimator, principal correlation estimator, principal component estimator, and generalized ridge and principal component estimator in terms of both the mean squared error (MSE) and Pitman closeness (PMC). Moreover, a numerical study on synthetic data has been performed to demonstrate the optimality of the proposed estimator.