Empirical Bayes procedures are commonly used based on the supposed asymptotic equivalence with fully Bayesian procedures, which, however, has not so far received full theoretical support in terms of uncertainty quantification. In this note, we provide some results on contraction rates of empirical Bayes posterior distributions which are illustrated in nonparametric curve estimation using Dirichlet process mixture models.
|Titolo:||On convergence rates of empirical Bayes procedures|
Scricciolo, Catia (Corresponding)
|Data di pubblicazione:||2014|
|Appare nelle tipologie:||04.01 Contributo in atti di convegno|