Confidence Sets for the Coefficients Vector of a Linear Regression Model with Nonspherical Disturbances
非球形扰动线性回归模型系数向量的置信集
In this present paper, considering a linear regression model with nonspherical disturbances, improved confidence sets for the regression coefficients vector are developed using the Stein rule estimators. We derive the large-sample approximations for the coverage probabilities and the expected volumes of the confidence sets based on the feasible generalized least-squares estimator and the Stein rule estimator and discuss their ranking.