Stringency in the emprical bayes problem
Dennis G. Haack · Communications in Statistics · 1975
The empirical Bayes approach to statistical decision problems is preasentad where the unknown prior distribution is estimated up to an equivalence −k. This approximation of the prior is than used in the decision process where an asymptotically stringent solution with respact to − k is sought. The equivalence relations − k are such that k becomes larger − k approachas equality and our decision rule becomes asymptotically stringent with respect to the unknown prior distribution.