Abstract
A new Liu type of estimator for the seemingly unrelated regression (SUR) models is proposed that may be used when estimating the parameters vector in the presence of multicollinearity if the it is suspected to belong to a linear subspace. The dispersion matrices and the mean squared error (MSE) are derived. The new estimator may have a lower MSE than the traditional estimators. It was shown using simulation techniques the new shrinkage estimator outperforms the commonly used estimators in the presence of multicollinearity.
DOI
10.22237/jmasm/1556669340
Recommended Citation
Månsson, K., Kibria, B. M. G., & Shukur, G. (2019). A new Liu type of estimator for the restricted SUR estimator. Journal of Modern Applied Statistical Methods, 18(1), eP2758. doi: 10.22237/jmasm/1556669340
Included in
Applied Statistics Commons, Social and Behavioral Sciences Commons, Statistical Theory Commons
