On Bayesian adaptation

S. Ghosal; J. Lember; A.W. van der Vaart

doi:10.1023/A:1025856016236

On Bayesian adaptation

S. Ghosal
, J. Lember
, A.W. van der Vaart

Mathematics

Research output: Contribution to Journal › Article › Academic › peer-review

Abstract

We show that Bayes estimators of an unknown density can adapt to unknown smoothness of the density. We combine prior distributions on each element of a list of log spline density models of different levels of regularity with a prior on the regularity levels to obtain a prior on the union of the models in the list. If the true density of the observations belongs to the model with a given regularity, then the posterior distribution concentrates near this true density at the rate corresponding to this regularity.

Original language	English
Pages (from-to)	165-175
Journal	Acta Applicandae Mathematicae
Volume	79
DOIs	https://doi.org/10.1023/A:1025856016236
Publication status	Published - 2003

Bibliographical note

MR2021886
Proceedings title: Proceedings of the Eighth Vilnius Conference on Probability Theory and Mathematical Statistics, Part II (2002)

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

SDG 14 Life Below Water

Access to Document

10.1023/A:1025856016236

Persistent URL (handle)

Persistent link

Cite this

@article{27fabc8fd06a4d5a8466e13cfc554cb6,

title = "On Bayesian adaptation",

abstract = "We show that Bayes estimators of an unknown density can adapt to unknown smoothness of the density. We combine prior distributions on each element of a list of log spline density models of different levels of regularity with a prior on the regularity levels to obtain a prior on the union of the models in the list. If the true density of the observations belongs to the model with a given regularity, then the posterior distribution concentrates near this true density at the rate corresponding to this regularity.",

author = "S. Ghosal and J. Lember and \{van der Vaart\}, A.W.",

note = "MR2021886 Proceedings title: Proceedings of the Eighth Vilnius Conference on Probability Theory and Mathematical Statistics, Part II (2002)",

year = "2003",

doi = "10.1023/A:1025856016236",

language = "English",

volume = "79",

pages = "165--175",

journal = "Acta Applicandae Mathematicae",

issn = "0167-8019",

publisher = "Springer Netherlands",

}

TY - JOUR

T1 - On Bayesian adaptation

AU - Ghosal, S.

AU - Lember, J.

AU - van der Vaart, A.W.

N1 - MR2021886 Proceedings title: Proceedings of the Eighth Vilnius Conference on Probability Theory and Mathematical Statistics, Part II (2002)

PY - 2003

Y1 - 2003

N2 - We show that Bayes estimators of an unknown density can adapt to unknown smoothness of the density. We combine prior distributions on each element of a list of log spline density models of different levels of regularity with a prior on the regularity levels to obtain a prior on the union of the models in the list. If the true density of the observations belongs to the model with a given regularity, then the posterior distribution concentrates near this true density at the rate corresponding to this regularity.

AB - We show that Bayes estimators of an unknown density can adapt to unknown smoothness of the density. We combine prior distributions on each element of a list of log spline density models of different levels of regularity with a prior on the regularity levels to obtain a prior on the union of the models in the list. If the true density of the observations belongs to the model with a given regularity, then the posterior distribution concentrates near this true density at the rate corresponding to this regularity.

U2 - 10.1023/A:1025856016236

DO - 10.1023/A:1025856016236

M3 - Article

SN - 0167-8019

VL - 79

SP - 165

EP - 175

JO - Acta Applicandae Mathematicae

JF - Acta Applicandae Mathematicae

ER -

On Bayesian adaptation

Abstract

Bibliographical note

UN SDGs

Access to Document

Persistent URL (handle)

Fingerprint

Cite this