Bayes/frequentist compromise decision rules for Gaussian sampling

Lynn E. Eberly, Thomas Louis

Research output: Contribution to journalArticle

Abstract

Bayesian methods have the potential to confer substantial advantages over frequentist when the assumed prior is approximately correct, but otherwise can perform poorly. Therefore, estimators and other inferences that strike a compromise between Bayes and frequentist optimality are attractive. To evaluate potential trade-offs, we study Bayes vs. frequentist risk under Gaussian sampling for families of point estimators and interval estimators. Bayes/frequentist compromises for interval estimation are more challenging than for point estimation, since performance involves an interplay between coverage and length. Each family allows 'purchasing' improved frequentist performance by allowing a small increase in Bayes risk over the Bayes rule. Any degree of increase can be specified, thus enabling greater or lesser trade-offs between Bayes and frequentist risk.

Original languageEnglish (US)
Pages (from-to)191-207
Number of pages17
JournalJournal of Statistical Planning and Inference
Volume121
Issue number2
DOIs
StatePublished - Apr 1 2004

Fingerprint

Bayes
Decision Rules
Sampling
Estimator
Trade-offs
Purchasing
Bayes Risk
Bayes Rule
Point Estimation
Interval Estimation
Bayesian Methods
Optimality
Coverage
Interval
Compromise
Decision rules
Evaluate
Family

Keywords

  • Bayes risk
  • Contamination
  • Frequentist risk
  • Mixture
  • Robustness
  • t-prior

ASJC Scopus subject areas

  • Statistics and Probability
  • Statistics, Probability and Uncertainty
  • Applied Mathematics

Cite this

Bayes/frequentist compromise decision rules for Gaussian sampling. / Eberly, Lynn E.; Louis, Thomas.

In: Journal of Statistical Planning and Inference, Vol. 121, No. 2, 01.04.2004, p. 191-207.

Research output: Contribution to journalArticle

@article{f8633a773e79462ca0a606d6347eeaec,
title = "Bayes/frequentist compromise decision rules for Gaussian sampling",
abstract = "Bayesian methods have the potential to confer substantial advantages over frequentist when the assumed prior is approximately correct, but otherwise can perform poorly. Therefore, estimators and other inferences that strike a compromise between Bayes and frequentist optimality are attractive. To evaluate potential trade-offs, we study Bayes vs. frequentist risk under Gaussian sampling for families of point estimators and interval estimators. Bayes/frequentist compromises for interval estimation are more challenging than for point estimation, since performance involves an interplay between coverage and length. Each family allows 'purchasing' improved frequentist performance by allowing a small increase in Bayes risk over the Bayes rule. Any degree of increase can be specified, thus enabling greater or lesser trade-offs between Bayes and frequentist risk.",
keywords = "Bayes risk, Contamination, Frequentist risk, Mixture, Robustness, t-prior",
author = "Eberly, {Lynn E.} and Thomas Louis",
year = "2004",
month = "4",
day = "1",
doi = "10.1016/S0378-3758(03)00134-4",
language = "English (US)",
volume = "121",
pages = "191--207",
journal = "Journal of Statistical Planning and Inference",
issn = "0378-3758",
publisher = "Elsevier",
number = "2",

}

TY - JOUR

T1 - Bayes/frequentist compromise decision rules for Gaussian sampling

AU - Eberly, Lynn E.

AU - Louis, Thomas

PY - 2004/4/1

Y1 - 2004/4/1

N2 - Bayesian methods have the potential to confer substantial advantages over frequentist when the assumed prior is approximately correct, but otherwise can perform poorly. Therefore, estimators and other inferences that strike a compromise between Bayes and frequentist optimality are attractive. To evaluate potential trade-offs, we study Bayes vs. frequentist risk under Gaussian sampling for families of point estimators and interval estimators. Bayes/frequentist compromises for interval estimation are more challenging than for point estimation, since performance involves an interplay between coverage and length. Each family allows 'purchasing' improved frequentist performance by allowing a small increase in Bayes risk over the Bayes rule. Any degree of increase can be specified, thus enabling greater or lesser trade-offs between Bayes and frequentist risk.

AB - Bayesian methods have the potential to confer substantial advantages over frequentist when the assumed prior is approximately correct, but otherwise can perform poorly. Therefore, estimators and other inferences that strike a compromise between Bayes and frequentist optimality are attractive. To evaluate potential trade-offs, we study Bayes vs. frequentist risk under Gaussian sampling for families of point estimators and interval estimators. Bayes/frequentist compromises for interval estimation are more challenging than for point estimation, since performance involves an interplay between coverage and length. Each family allows 'purchasing' improved frequentist performance by allowing a small increase in Bayes risk over the Bayes rule. Any degree of increase can be specified, thus enabling greater or lesser trade-offs between Bayes and frequentist risk.

KW - Bayes risk

KW - Contamination

KW - Frequentist risk

KW - Mixture

KW - Robustness

KW - t-prior

UR - http://www.scopus.com/inward/record.url?scp=0742289091&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0742289091&partnerID=8YFLogxK

U2 - 10.1016/S0378-3758(03)00134-4

DO - 10.1016/S0378-3758(03)00134-4

M3 - Article

AN - SCOPUS:0742289091

VL - 121

SP - 191

EP - 207

JO - Journal of Statistical Planning and Inference

JF - Journal of Statistical Planning and Inference

SN - 0378-3758

IS - 2

ER -