Solmaz Seifollahi - Department of Statistics, University of Tabriz, Tabriz, Iran
Hossein Bevrani - Department of Statistics, University of Tabriz, Tabriz, Iran
Kaniav Kamary - CentraleSuplec, Laboratoire MICS, Paris, France

This paper focus on estimating of the multivariate multiple linear models (MMLMs)when multicollinearity presents in the design matrix. Among the estimators proposedto tackle the problem of multicollinearity, the ridge estimator is popular. We proposethe ridge-type estimators of the coefficient matrix when the multicollinearity existsalong with the prior information about the coefficients. A simulation study is utilizedto compare the relative efficiency of proposed estimators and finally, we analyze a realdataset by the proposed estimators to show the usefulness of the proposed estimators.

Ordinary ridge estimator, restricted ridge-type estimator, James-Steinridge-type estimator, preliminary test ridge-type estimator.