Answers are grounded in this paper and cite their sources.
Empirical Bayes estimators, asymptotically optimal with rates, are proposed. In the component problem there is a pair (X, ω) of real valued random variables. The Lebesgue density of X, conditional on ω, is of the form u(x)C(ω)eω x. Based on a realization of X, the problem is squared error loss estimation of ω. Let G be a prior distribution on ω, and $R(G)$ be the Bayes optimal risk wrt G. Using (X₁, ⋯, Xₙ), the observations in the past n such problems, mean square consistent estimators of the derivative of log (∫ C(ω)eω x dG(ω)) are proposed. Then these statistics and the present observation X are used to exhibit estimators ψₙ for the present problem whose risks Rₙ converge to the Bayes optimal risk $R(G)$ as n → ∞. In particular, with no assumption on the smoothness or on the form of u, a ψₙ for each γ in [0, 2) is exhibited. Sufficient conditions are given under which c₁n-4/(4+3γ) Rₙ - R(G) c₂n-2γ/(4+3γ), where c₁ and c₂ are positive constants. The rhs inequality holds uniformly in G with support in a bounded interval of the real line, while the other holds for a G degenerate at a point and for all n sufficiently large. (Thus with γ close to 2, ψₙ achieves almost the exact rate.) Examples of families, including one whose u function has infinitely many discontinuities, are given where conditions for the above inequalities are satisfied for γ arbitrarily close to 2.
No takes yet. Share an insight, caveat, or question.
R. S. Singh (1976) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: