Limiting the Risk of Bayes and Empirical Bayes Estimators, Part II: The Empirical Bayes Case.
Introduces modifications of Bayes and empirical Bayes estimators which do nearly as well as their prototypes when the assumptions about the a priori information are valid, but which are not subject to the gross errors of their prototypes when a priori information is misleading. This problem is discussed in the context of simultaneous estimation of the means of several normal distributions, each of the means being taken from the same unknown a priori distribution. Several tables and figures contain values needed for application of the procedures. Some simple applications illustrate the methods. 80 pp. Ref. (Author)