A modified two-parameter estimator in linear regression

Abstract

In this article, a modified two-parameter estimator is introduced for the vector of parameters in the linear regression model when data exists with multicollinearity. The properties of the proposed estimator are discussed and the performance in terms of the matrix mean square error criterion over the ordinary least squares (OLS) estimator, a new two-parameter estimator (NTP), an almost unbiased two-parameter estimator (AUTP) and other well known estimators reviewed in this article is investigated. A numerical example and simulation study are finally conducted to illustrate the superiority of the proposed estimator.