A Bayes but Not Classically Sufficient Statistic
D. Blackwell, R. V. Ramamoorthi
The Annals of Statistics · 1982 · Volume 10 · Issue 3
Abstract
In a Borel setting, every classically sufficient statistic is Bayes sufficient, but not vice versa. The example is a hypothesis testing problem in which Bayesians, but not classicists, can achieve zero error probabilities.
Keywords
62B20, 62F15, Bayes, Sufficient statistic
Enable JavaScript for citation tools and additional paper interactions.