Feasible Stein-Type and Preliminary Test Estimations in the System Regression Model

In a system of regression models, finding feasible shrinkage is demanding since the covariance structure unknown and cannot be ignored. On other hand, specifying sub-space restrictions for adequate vital. This study proposes estimation strategies where restriction obtained from LASSO. Therefore, some LASSO-based Stein-type estimators are introduced, their asymptotic performance studied. Extensive Monte Carlo simulation real-data experiment support superior proposed compared to generalized least-squared estimator.