Answers are grounded in this paper and cite their sources.
Let Sp × p have a Wishart distribution with scale matrix Σ and k degrees of freedom. Estimators of Σ are given for each of the loss functions L₁(Σ̂, Σ) = tr (Σ̂Σ⁻¹) - log (Σ̂Σ⁻¹) - p and L₂(Σ̂, Σ) = tr (Σ̂Σ⁻¹ - I)². The obvious estimators of Σ are the scalar multiples of S, i.e., aS where 0 < a 1/k. (Recall that (1/k)S is unbiased.) For each problem (Σ, Σ̂, Lᵢ), i = 1, 2, we provide empirical Bayes estimators which dominate aS by a substantial amount. It is seen that the uniform reduction in the risk function determined by L₂ is at least 100(p + 1)/(k + p + 1)%. Dominance results for L₁ and L₂ were first given by James and Stein.
No takes yet. Share an insight, caveat, or question.
L. R. Haff (1980) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: