Abstract
We study full Bayesian procedures for high-dimensional linear regression. We adopt data-dependent empirical priors introduced in Martin et al. (Bernoulli 23(3):1822–1847, 2017). In their paper, these priors have nice posterior contraction properties and are easy to compute. Our paper extend their theoretical results to the case of unknown error variance . Under proper sparsity assumption, we achieve model selection consistency, posterior contraction rates as well as Bernstein von-Mises theorem by analyzing multivariate t-distribution.
Similar content being viewed by others
References
Arias-Castro E, Lounici K (2014) Estimation and variable selection with exponential weights. Electron J Stat 8(1):328–354
Bondell H, Reich B (2012) Consistent high-dimensional Bayesian variable selection via penalized credible regions. J Am Stat Assoc 107:1601–1624
Cao X, Khare K, Ghosh M (2020) High-dimensional posterior consistency for hierarchical non-local priors in regression. Bayesian Anal 15(1):241–262
Castillo I, Schmidt-Hieber J, van der vaart A (2015) Bayesian linear regression with sparse priors. Ann Stat 43(5):1986–2018
de Geer SAV, Buhlmann P (2009) On the conditions used to prove oracle results for the lasso. Electron J Stat 3:1360–1392
Grünwald P, van Ommen T (2017) Inconsistency of Bayesian inference for misspecified linear models, and a proposal for repairing it. Bayesian Anal 12:1069–1103
Holmes CC, Walker SG (2017) Assigning a value to a power likelihood in a general Bayesian model. Biometrika 104(2):497–503
Hong L, Kuffner TA, Martin R (2018) On overfitting and post-selection uncertainty assessments. Biometrika 105(1):221–224
Ishwaran H, Rao J (2005) Spike and slab variable selection: frequentist and Bayesian strategies. Ann Stat 33:730–773
Johnson VE, Rossell D (2012) Bayesian model selection in high-dimensional settings. J Am Stat Assoc 107:649–660
Martin R, Ning B (2020) Empirical priors and coverage of posterior credible sets in a sparse normal mean model. Indian J Stat 82–A:477–498
Martin R, Tang Y (2020) Empirical priors for prediction in sparse high-dimensional linear regression. J Mach Learn Res 21:1–30
Martin R, Walker S (2014) Asymptotically minimax empirical Bayes estimation of a sparse normal mean vector. Electron J Stat 8(2):2188–2206
Martin R, Walker S (2019) Data-dependent priors and their posterior concentration rates. Electron J Stat 13:3049–3081
Martin R, Mess R, Walker S (2017) Empirical bayes posterior concentration in sparse high-dimensional linear models. Bernoulli 23(3):1822–1847
Narisetty N, He X (2014) Bayesian variable selection with shrinking and diffusing priors. Ann Stat 42:789–817
Shin M, Bhattacharya A, Johnson VE (2019) Functional horseshoe priors for subspace shrinkage. J Am Stat Assoc 115(532):1784–1797
Syring M, Martin R (2019) Calibrating general posterior credible regions. Biometrika 106:479–486
Wainwright MJ (2009) Sharp thresholds for high-dimensional and noisy sparsity recovery using l1 -constrained quadratic programming (lasso). IEEE Trans 55:2183–2202
Zou H (2006) The adaptive lasso and its oracle properties. J Am Stat Assoc 101:1418–1429
Acknowledgements
Thanks are due to two reviewers for many helpful suggestions.
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Appendix A: Tail bounds for the Chi squared distribution
Appendix A: Tail bounds for the Chi squared distribution
Lemma 2
For any \(a>0\), we have
Proof: See Lemma 4.1 of Cao et al. (2020).
Lemma 3
(i) For any \(c>0\), we have
(ii) for \(\omega <1\) then
where \(c_1>0\) is a constant.
(iii) For any \(c>0\), \(P(\chi ^2_p(\lambda )-p \le -c) \le \exp \left( -\frac{c^2}{4p}\right) \).
Proof: See (Cao et al. 2020) Lemma 4.2 for (i) and for (ii) on (Shin et al. 2019). (iii) follows from the fact if \(U \sim \chi ^2_p(\lambda )\), then \(P(U>c)\) is strictly increasing in \(\lambda \) for fixed p and \(c>0\). Hence
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Fang, X., Ghosh, M. High-dimensional properties for empirical priors in linear regression with unknown error variance. Stat Papers 65, 237–262 (2024). https://doi.org/10.1007/s00362-022-01390-0
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00362-022-01390-0