A NOTE ON GOOD RIDGE ESTIMATORS BASED ON PRIOR INFORMATION

Theil (1963) suggested a mixed estimator when there is prior information used in regression analysis. It implies that the investigator not only estimates a parameter based on the sample data, but also combines the prior information by formulating it in terms of unbiased estimates and their covariance matrix. In particular, he assumes that the prior unbiased estimators r are q linear combinations of the parameters. The mixed regression model is