半相依回归方程组两步估计的优良性
OPTIMALITY OF TWO-STAGE ESTIMATORS OF SEEMINGLY UNRELATED REGRESSION EQUATIONS
-
Abstract: Zellner’s two-stage estimators of a kind of regression equation systems are discussed. Without using the assumptions imposed by Zellner and Revankar on the relations of the design matrices,we obtain the two-stage estimator(TSE)and prove that the TSE is better than the Least Square Estimator.