Bayesian model criticism using uniform parametrization checks

Christian T. Covington, Jeffrey W. Miller · 2025

Read Bayesian model criticism using uniform parametrization checks by Christian T. Covington, Jeffrey W. Miller (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2503-18261.txt.

Description

Christian Covington and Jeffrey Miller propose a new way to check Bayesian models in this two thousand twenty-five paper. In practice, models are often wrong in part. A good check should show not just that a model fails, but where it fails. Their method rests on a neat fact. If data truly came from the assumed model, then a draw from the posterior should look like a draw from the prior.

They rewrite every random part of the model in terms of plain uniform values. These u-values can be grouped and tested in many ways. A group that strays from uniform points at the flawed part of the model. The checks are cheap to run and rest on firm statistics. The paper gives Bayesian analysts a sharper tool for model criticism.

Catalog record

CREATORChristian T. Covington, Jeffrey W. Miller
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE95 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2503-18261

Browse the full library · About the Tower of Babel