Papers
Topics
Authors
Recent
Search
2000 character limit reached

The distribution of Bayes' ratio

Published 31 Mar 2024 in astro-ph.CO, math.ST, and stat.TH | (2404.00744v3)

Abstract: 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.

Definition Search Book Streamline Icon: https://streamlinehq.com
References (38)
  1. Constraints on cosmic curvature from cosmic chronometer and quasar observations. arXiv e-prints, art. arXiv:2303.15401, March 2023. doi: 10.48550/arXiv.2303.15401.
  2. Restoring cosmological concordance with early dark energy and massive neutrinos? Mon. Not. Roy. Astron. Soc., 520(3):3688–3695, 2023. doi: 10.1093/mnras/stad317.
  3. Constraining cosmology with the Gaia-unWISE Quasar Catalog and CMB lensing: structure growth. JCAP, 11:043, 2023. doi: 10.1088/1475-7516/2023/11/043.
  4. Observational constraints on early dark energy. arXiv e-prints, art. arXiv:2310.19899, October 2023. doi: 10.48550/arXiv.2310.19899.
  5. Constraining f(R) gravity with cross-correlation of galaxies and cosmic microwave background lensing. arXiv e-prints, art. arXiv:2311.09936, November 2023. doi: 10.48550/arXiv.2311.09936.
  6. Latest Data Constraint of Some Parameterized Dark Energy Models. Chin. Phys. Lett., 40(1):019801, 2023. doi: 10.1088/0256-307X/40/1/019801.
  7. Bayesian and frequentist investigation of prior effects in EFT of LSS analyses of full-shape BOSS and eBOSS data. Phys. Rev. D, 108(12):123514, 2023. doi: 10.1103/PhysRevD.108.123514.
  8. Constraining LLTB models with galaxy cluster counts from next generation surveys. arXiv e-prints, art. arXiv:2309.17151, September 2023. doi: 10.48550/arXiv.2309.17151.
  9. MCMC Marginalisation Bias and ΛΛ\Lambdaroman_ΛCDM tensions. 7 2023.
  10. Cluster counts - III. ΛΛ\Lambdaroman_ΛCDM extensions and the cluster tension. Astron. Astrophys., 666:A34, 2022. doi: 10.1051/0004-6361/202142115.
  11. Bayesian evidence for a cosmological constant using new high-redshift supernova data. Monthly Notices of the Royal Astronomical Society, 379(1):169–175, July 2007. ISSN 1365-2966. doi: 10.1111/j.1365-2966.2007.11924.x. URL http://dx.doi.org/10.1111/j.1365-2966.2007.11924.x.
  12. The ΛΛ\Lambdaroman_ΛCDM Model in the Lead—A Bayesian Cosmological Model Comparison. Astrophys. J. , 675(1):1–7, March 2008. doi: 10.1086/526333.
  13. Comparison of cosmological models using Bayesian theory. Phys. Rev. D, 65(4):043506, February 2002. doi: 10.1103/PhysRevD.65.043506.
  14. Daniel L. Shafer. Robust model comparison disfavors power law cosmology. Phys. Rev. D, 91(10):103516, 2015. doi: 10.1103/PhysRevD.91.103516.
  15. No evidence for extensions to the standard cosmological model. Phys. Rev. Lett., 119(10):101301, 2017. doi: 10.1103/PhysRevLett.119.101301.
  16. H. Jeffreys. Theory of probability. Oxford Classics series (reprinted 1998). Oxford University Press, Oxford, UK, 3rd edn edition, 1961.
  17. Harold Jeffreys’s Theory of Probability Revisited. arXiv e-prints, art. arXiv:0804.3173, April 2008. doi: 10.48550/arXiv.0804.3173.
  18. Bayes Factors. J. Am. Statist. Assoc., 90(430):773–795, 1995. doi: 10.1080/01621459.1995.10476572.
  19. G. Efstathiou. Limitations of Bayesian Evidence Applied to Cosmology. Mon. Not. Roy. Astron. Soc., 388:1314, 2008. doi: 10.1111/j.1365-2966.2008.13498.x.
  20. Is the Jeffreys’ scale a reliable tool for Bayesian model comparison in cosmology? JCAP, 08:036, 2013. doi: 10.1088/1475-7516/2013/08/036.
  21. The power of Bayesian evidence in astronomy. Mon. Not. Roy. Astron. Soc., 413:2895, 2011. doi: 10.1111/j.1365-2966.2011.18361.x.
  22. Comparison between different methods of model selection in cosmology. Eur. Phys. J. Plus, 136(2):219, 2021. doi: 10.1140/epjp/s13360-021-01200-w.
  23. Bayesian model averaging: a tutorial (with comments by M. Clyde, David Draper and E. I. George, and a rejoinder by the authors. Statistical Science, 14(4):382 – 417, 1999. doi: 10.1214/ss/1009212519. URL https://doi.org/10.1214/ss/1009212519.
  24. Bayesian Model Averaging in Astrophysics: A Review. Statist. Anal. Data Mining, 6:3–14, 2013. doi: 10.1002/sam.11179.
  25. Measuring the effective complexity of cosmological models. Phys. Rev., D74:023503, 2006. doi: 10.1103/PhysRevD.74.023503.
  26. Concordance and Discordance in Cosmology. Phys. Rev. D, 99(4):043506, 2019. doi: 10.1103/PhysRevD.99.043506.
  27. Extreme data compression for Bayesian model comparison. jcap, 2023(11):048, November 2023. doi: 10.1088/1475-7516/2023/11/048.
  28. Bayesian vs frequentist: comparing Bayesian model selection with a frequentist approach using the iterative smoothing method. JCAP, 03(03):047, 2022. doi: 10.1088/1475-7516/2022/03/047.
  29. I. J. Good. Saddle-point Methods for the Multinomial Distribution. The Annals of Mathematical Statistics, 28(4):861 – 881, 1957. doi: 10.1214/aoms/1177706790. URL https://doi.org/10.1214/aoms/1177706790.
  30. I. J. Good. The bayes/non-bayes compromise: A brief review. Journal of the American Statistical Association, 87(419):597–606, 1992. ISSN 01621459. URL http://www.jstor.org/stable/2290192.
  31. Internal Robustness: systematic search for systematic bias in SN Ia data. Mon.Not.Roy.Astron.Soc., 430:1867–1879, 2013. doi: 10.1093/mnras/stt008.
  32. Kevin Wolz. Bayesian and frequentist methods to investigate tensions between data sets. Msc thesis, University of Heidelberg, Heidelberg, Germany, October 2018.
  33. in prep. 2024.
  34. On the distribution of Bayesian evidence. Mon. Not. Roy. Astron. Soc., 515(1):293–301, 2022. doi: 10.1093/mnras/stac1851.
  35. Objective Bayesian analysis of neutrino masses and hierarchy. jcap, 2018(4):047, April 2018. doi: 10.1088/1475-7516/2018/04/047.
  36. Amanullah et al. Spectra and Hubble Space Telescope Light Curves of Six Type Ia Supernovae at 0.511 ¡ z ¡ 1.12 and the Union2 Compilation. Ap.J., 716:712–738, June 2010. doi: 10.1088/0004-637X/716/1/712.
  37. J. P. IMHOF. Computing the distribution of quadratic forms in normal variables. Biometrika, 48(3-4):419–426, 12 1961. ISSN 0006-3444. doi: 10.1093/biomet/48.3-4.419. URL https://doi.org/10.1093/biomet/48.3-4.419.
  38. Pierre Duchesne and Pierre Lafaye De Micheaux. Computing the distribution of quadratic forms: Further comparisons between the liu-tang-zhang approximation and exact methods. Computational Statistics & Data Analysis, 54(4):858–862, 2010. ISSN 0167-9473. doi: https://doi.org/10.1016/j.csda.2009.11.025. URL https://www.sciencedirect.com/science/article/pii/S0167947309004381.
Citations (3)
View on Semantic Scholar

Summary

No one has generated a summary of this paper yet.

Sign Up to Summarize

Paper to Video (Beta)

No one has generated a video about this paper yet.

Sign Up to Generate All Videos Subscribe on YouTube

Whiteboard

No one has generated a whiteboard explanation for this paper yet.

Sign Up to Generate

Open Problems

We haven't generated a list of open problems mentioned in this paper yet.

Generate Now

Continue Learning

We haven't generated follow-up questions for this paper yet.

Generate Now

Collections

Sign up for free to add this paper to one or more collections.

Sign Up

Tweets

Sign up for free to view the 2 tweets with 3 likes about this paper.

Sign Up for Free