Guardado en:
Detalles Bibliográficos
Autores principales: Amendola, Luca, Patel, Vrund, Sakr, Ziad, Sellentin, Elena, Wolz, Kevin
Formato: Preprint
Publicado: 2024
Materias:
Acceso en línea:https://arxiv.org/abs/2404.00744
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866912153885212672
author Amendola, Luca
Patel, Vrund
Sakr, Ziad
Sellentin, Elena
Wolz, Kevin
author_facet Amendola, Luca
Patel, Vrund
Sakr, Ziad
Sellentin, Elena
Wolz, Kevin
contents The ratio of Bayesian evidences is a popular tool in cosmology to compare different models. There are however several issues with this method: Bayes' ratio depends on the prior even in the limit of non-informative priors, and Jeffrey's scale, used to assess the test, is arbitrary. Moreover, the standard use of Bayes' ratio is often criticized for being unable to reject models. In this paper, we address these shortcoming by promoting evidences and evidence ratios to frequentist statistics and deriving their sampling distributions. By comparing the evidence ratios to their sampling distributions, poor fitting models can now be rejected. Our method additionally does not depend on the prior in the limit of very weak priors, thereby safeguarding the experimenter against premature rejection of a theory with a uninformative prior, and replaces the arbitrary Jeffrey's scale by probability thresholds for rejection. We provide analytical solutions for some simplified cases (Gaussian data, linear parameters, and nested models), and we apply the method to cosmological supernovae Ia data. We dub our method the FB method, for Frequentist-Bayesian.
format Preprint
id arxiv_https___arxiv_org_abs_2404_00744
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle The distribution of Bayes' ratio
Amendola, Luca
Patel, Vrund
Sakr, Ziad
Sellentin, Elena
Wolz, Kevin
Cosmology and Nongalactic Astrophysics
Statistics Theory
The ratio of Bayesian evidences is a popular tool in cosmology to compare different models. There are however several issues with this method: Bayes' ratio depends on the prior even in the limit of non-informative priors, and Jeffrey's scale, used to assess the test, is arbitrary. Moreover, the standard use of Bayes' ratio is often criticized for being unable to reject models. In this paper, we address these shortcoming by promoting evidences and evidence ratios to frequentist statistics and deriving their sampling distributions. By comparing the evidence ratios to their sampling distributions, poor fitting models can now be rejected. Our method additionally does not depend on the prior in the limit of very weak priors, thereby safeguarding the experimenter against premature rejection of a theory with a uninformative prior, and replaces the arbitrary Jeffrey's scale by probability thresholds for rejection. We provide analytical solutions for some simplified cases (Gaussian data, linear parameters, and nested models), and we apply the method to cosmological supernovae Ia data. We dub our method the FB method, for Frequentist-Bayesian.
title The distribution of Bayes' ratio
topic Cosmology and Nongalactic Astrophysics
Statistics Theory
url https://arxiv.org/abs/2404.00744