Improving the Stein-Rule Estimator of Each Individual Regression Coefficient Using the Stein Varience Estimator

摘要: A substantial body of literature on regression has focused estimators which are biased but more precise than the ordinary least squares (OLS) estimator. As is discussed in Paelinck and Klaasseen (1979), estimation parameters important spatial econometric models, precision often measured by mean square error (MSE). It well known that when three or coefficients estimated simultaneously, Stein-rule (SR) estimator proposed Stein (1955) James (1961) dominates OLS terms MSE. [Exactly speaking, predictive MSE.] Further, positive-part (PSR) SR [See, for example, Judge Yancey (1986).