Calibrating general posterior credible regions. (10th December 2018)

• Record Type: Journal Article
• Authors: Syring, Nicholas; Martin, Ryan
• Is Part Of: Biometrika. Volume 106:Number 2(2019:Jun.)
• Page Start: 479
• Page End: 486
• Publication Date: 2018-12-10
• Subjects: Bootstrap -- Coverage probability -- Gibbs posterior distribution -- Model misspecification -- Stochastic approximation; Biometry -- Periodicals; 570.1519505
• DOI: 10.1093/biomet/asy054
• Languages: English
• ISSNs: 0006-3444

Abstract

Summary: Calibration of credible regions derived from under- or misspecified models is an important and challenging problem. In this paper, we introduce a scalar tuning parameter that controls the posterior distribution spread, and develop a Monte Carlo algorithm that sets this parameter so that the corresponding credible region achieves the nominal frequentist coverage probability.