This paper considers the predictive efficiency of some improved estimators when there are exact prior restrictions on the coefficients of the linear regression model. A comparative study of various improved estimators in terms of their respective risks is also described.
展开▼