The generalized new two-type parameter estimator in linear regression model

Abstract

In this paper, a new two-type parameter estimator is proposed. This estimator is a generalization of the new two parameter (NTP) estimator introduced by Yang and Chang, which includes the ordinary least squares (OLS), the generalized ridge (GR) and the generalized Liu (GL) estimators, as special cases. Here, the performance of this new estimator is, theoretically, investigated over the OLS, the GR, the GL and the NTP estimators in terms of mean squared error matrix criterion. Furthermore, the estimation of the biasing parameters is obtained to minimize the scalar mean squared error. In addition, a complementary algorithm is proposed for the estimator presented by Yang and Chang. As well, a numerical example is given and a simulation study is done.

Keywords:

• Generalized Liu estimator
• Generalized ridge estimator
• Lagrange method
• Mean squared error
• Two-parameter estimator