[Accepted Manuscript] Tractable Bayesian variable selection: beyond normality


Rossell, D.; Rubio, F.J.; (2017) [Accepted Manuscript] Tractable Bayesian variable selection: beyond normality. Journal of the American Statistical Association. ISSN 0162-1459 DOI: https://doi.org/10.1080/01621459.2017.1371025

WarningThere is a more recent version of this item available.
[img] Text - Accepted Version
Restricted to Repository staff only until 14 March 2019.
License:

Download (829kB) | Request a copy
[img] Text - Supplemental Material
Restricted to Repository staff only until 14 March 2019.
License:

Download (1MB) | Request a copy

Abstract

AbstractBayesian variable selection often assumes normality, but the effects of model misspecification are not sufficiently understood. There are sound reasons behind this assumption, particularly for large p: ease of interpretation, analytical and computational convenience. More flexible frameworks exist, including semi- or non-parametric models, often at the cost of some tractability. We propose a simple extension that allows for skewness and thicker-than-normal tails but preserves tractability. It leads to easy interpretation and a log-concave likelihood that facilitates optimization and integration. We characterize asymptotically parameter estimation and Bayes factor rates, under certain model misspecification. Under suitable conditions misspecified Bayes factors induce sparsity at the same rates than under the correct model. However, the rates to detect signal change by an exponential factor, often reducing sensitivity. These deficiencies can be ameliorated by inferring the error distribution, a simple strategy that can improve inference substantially. Our work focuses on the likelihood and can be combined with any likelihood penalty or prior, but here we focus on non-local priors to induce extra sparsity and ameliorate finite-sample effects caused by misspecification. We show the importance of considering the likelihood rather than solely the prior, for Bayesian variable selection. The methodology is in R package ?mombf?.

Item Type: Article
Faculty and Department: Faculty of Epidemiology and Population Health > Dept of Non-Communicable Disease Epidemiology
URI: http://researchonline.lshtm.ac.uk/id/eprint/4645529

Available Versions of this Item

Statistics


Download activity - last 12 months
Downloads since deposit
2Downloads
60Hits
Accesses by country - last 12 months
Accesses by referrer - last 12 months
Impact and interest
Additional statistics for this record are available via IRStats2

Actions (login required)

Edit Item Edit Item