Honest Bayesian confidence sets for the L2-norm

B.T. Szabó, A.W. Vaart, van der, J.H. Zanten, van

Onderzoeksoutput: Bijdrage aan tijdschriftTijdschriftartikelAcademicpeer review

15 Citaten (Scopus)


We investigate the problem of constructing Bayesian credible sets that are honest and adaptive for the L^2-loss over a scale of Sobolev classes with regularity ranging between [D,2D], for some given D in the context of the signal-in-white-noise model. We consider a scale of prior distributions indexed by a regularity hyper-parameter and choose the hyper-parameter both by marginal likelihood empirical Bayes and by hierarchical Bayes method, respectively. Next we consider a ball centred around the corresponding posterior mean with prescribed posterior probability. We show by theory and examples that both the empirical Bayes and the hierarchical Bayes credible sets give misleading, overconfident uncertainty quantification for certain oddly behaving truth. Then we construct a new empirical Bayes method based on risk estimation, which provides the correct uncertainty quantification and optimal size. Keywords: Credible sets; Coverage; Uncertainty quantification
Originele taal-2Engels
Pagina's (van-tot)36-51
Aantal pagina's16
TijdschriftJournal of Statistical Planning and Inference
StatusGepubliceerd - 2015


Duik in de onderzoeksthema's van 'Honest Bayesian confidence sets for the L2-norm'. Samen vormen ze een unieke vingerafdruk.

Citeer dit