Bayesian sensitivity analysis with the Fisher–Rao metric

We propose a geometric framework to assess sensitivity of Bayesian procedures to modelling assumptions based on the nonparametric Fisher–Rao metric. While the framework is general, the focus of this article is on assessing local and global robustness in Bayesian procedures with respect to perturbati...

Full description

Bibliographic Details
Main Authors: Kurtek, Sebastian, Bharath, Karthik
Format: Article
Published: Oxford University Press 2015
Subjects:
Online Access:https://eprints.nottingham.ac.uk/40807/
_version_ 1848796138257252352
author Kurtek, Sebastian
Bharath, Karthik
author_facet Kurtek, Sebastian
Bharath, Karthik
author_sort Kurtek, Sebastian
building Nottingham Research Data Repository
collection Online Access
description We propose a geometric framework to assess sensitivity of Bayesian procedures to modelling assumptions based on the nonparametric Fisher–Rao metric. While the framework is general, the focus of this article is on assessing local and global robustness in Bayesian procedures with respect to perturbations of the likelihood and prior, and on the identification of influential observations. The approach is based on a square-root representation of densities, which enables analytical computation of geodesic paths and distances, facilitating the definition of naturally calibrated local and global discrepancy measures. An important feature of our approach is the definition of a geometric ϵ-contamination class of sampling distributions and priors via intrinsic analysis on the space of probability density functions. We demonstrate the applicability of our framework to generalized mixed-effects models and to directional and shape data.
first_indexed 2025-11-14T19:43:13Z
format Article
id nottingham-40807
institution University of Nottingham Malaysia Campus
institution_category Local University
last_indexed 2025-11-14T19:43:13Z
publishDate 2015
publisher Oxford University Press
recordtype eprints
repository_type Digital Repository
spelling nottingham-408072020-05-04T17:13:02Z https://eprints.nottingham.ac.uk/40807/ Bayesian sensitivity analysis with the Fisher–Rao metric Kurtek, Sebastian Bharath, Karthik We propose a geometric framework to assess sensitivity of Bayesian procedures to modelling assumptions based on the nonparametric Fisher–Rao metric. While the framework is general, the focus of this article is on assessing local and global robustness in Bayesian procedures with respect to perturbations of the likelihood and prior, and on the identification of influential observations. The approach is based on a square-root representation of densities, which enables analytical computation of geodesic paths and distances, facilitating the definition of naturally calibrated local and global discrepancy measures. An important feature of our approach is the definition of a geometric ϵ-contamination class of sampling distributions and priors via intrinsic analysis on the space of probability density functions. We demonstrate the applicability of our framework to generalized mixed-effects models and to directional and shape data. Oxford University Press 2015-07-15 Article PeerReviewed Kurtek, Sebastian and Bharath, Karthik (2015) Bayesian sensitivity analysis with the Fisher–Rao metric. Biometrika, 102 (3). pp. 601-616. ISSN 0006-3444 Fisher–Rao metric Geodesic Geometric ϵ-contamination Influence analysis Riemannian manifold https://academic.oup.com/biomet/article-lookup/doi/10.1093/biomet/asv026 doi:10.1093/biomet/asv026 doi:10.1093/biomet/asv026
spellingShingle Fisher–Rao metric
Geodesic
Geometric ϵ-contamination
Influence analysis
Riemannian manifold
Kurtek, Sebastian
Bharath, Karthik
Bayesian sensitivity analysis with the Fisher–Rao metric
title Bayesian sensitivity analysis with the Fisher–Rao metric
title_full Bayesian sensitivity analysis with the Fisher–Rao metric
title_fullStr Bayesian sensitivity analysis with the Fisher–Rao metric
title_full_unstemmed Bayesian sensitivity analysis with the Fisher–Rao metric
title_short Bayesian sensitivity analysis with the Fisher–Rao metric
title_sort bayesian sensitivity analysis with the fisher–rao metric
topic Fisher–Rao metric
Geodesic
Geometric ϵ-contamination
Influence analysis
Riemannian manifold
url https://eprints.nottingham.ac.uk/40807/
https://eprints.nottingham.ac.uk/40807/
https://eprints.nottingham.ac.uk/40807/