Matrix mean squared error comparisons of some biased estimators with two biasing parameters

KURNAZ F. S., Akay K. U.

COMMUNICATIONS IN STATISTICS-THEORY AND METHODS, vol.47, no.8, pp.2022-2035, 2018 (SCI-Expanded) identifier identifier

  • Publication Type: Article / Article
  • Volume: 47 Issue: 8
  • Publication Date: 2018
  • Doi Number: 10.1080/03610926.2017.1335415
  • Journal Name: COMMUNICATIONS IN STATISTICS-THEORY AND METHODS
  • Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED), Scopus
  • Page Numbers: pp.2022-2035
  • Keywords: Biased estimators, k - d class estimator, Liu-type estimator, Matrix mean squared error, Multicollinearity, RIDGE-REGRESSION, LINEAR-REGRESSION
  • Yıldız Technical University Affiliated: Yes

Abstract

To deal with multicollinearity problem, the biased estimators with two biasing parameters have recently attracted much research interest. The aim of this article is to compare one of the last proposals given by Yang and Chang (2010) with Liu-type estimator (Liu 2003) and k - d class estimator (Sakallioglu and Kaciranlar 2008) under the matrix mean squared error criterion. As well as giving these comparisons theoretically, we support the results with the extended simulation studies and real data example, which show the advantages of the proposal given by Yang and Chang (2010) over the other proposals with increasing multicollinearity level.