Bayesian model averaging is a practical method for dealing with uncertainty due to model specification. Use of this technique requires the estimation of model probability weights. Here, we revisit the derivation of estimators for these model weights. Use of the Kullback-Leibler divergence as a starting point leads naturally to a number of alternative information criteria suitable for Bayesian model weight estimation. We explore three such criteria, known to the statistics literature before, in detail: a Bayesian analog of the Akaike information criterion which we call the BAIC, the Bayesian predictive information criterion, and the posterior predictive information criterion (PPIC). We compare the use of these information criteria in numerical analysis problems common in lattice field theory calculations. We find that the PPIC has the most appealing theoretical properties and can give the best performance in terms of model-averaging uncertainty, particularly in the presence of noisy data, while the BAIC is a simple and reliable alternative.
Neil, Ethan T. and Sitison, Jacob W.. "Improved information criteria for Bayesian model averaging in lattice field theory." Physical Review. D., vol. 109, no. 1, Jan. 2024. https://doi.org/10.1103/physrevd.109.014510
Neil, Ethan T., & Sitison, Jacob W. (2024). Improved information criteria for Bayesian model averaging in lattice field theory. Physical Review. D., 109(1). https://doi.org/10.1103/physrevd.109.014510
Neil, Ethan T., and Sitison, Jacob W., "Improved information criteria for Bayesian model averaging in lattice field theory," Physical Review. D. 109, no. 1 (2024), https://doi.org/10.1103/physrevd.109.014510
@article{osti_2283673,
author = {Neil, Ethan T. and Sitison, Jacob W.},
title = {Improved information criteria for Bayesian model averaging in lattice field theory},
annote = {Bayesian model averaging is a practical method for dealing with uncertainty due to model specification. Use of this technique requires the estimation of model probability weights. Here, we revisit the derivation of estimators for these model weights. Use of the Kullback-Leibler divergence as a starting point leads naturally to a number of alternative information criteria suitable for Bayesian model weight estimation. We explore three such criteria, known to the statistics literature before, in detail: a Bayesian analog of the Akaike information criterion which we call the BAIC, the Bayesian predictive information criterion, and the posterior predictive information criterion (PPIC). We compare the use of these information criteria in numerical analysis problems common in lattice field theory calculations. We find that the PPIC has the most appealing theoretical properties and can give the best performance in terms of model-averaging uncertainty, particularly in the presence of noisy data, while the BAIC is a simple and reliable alternative.},
doi = {10.1103/physrevd.109.014510},
url = {https://www.osti.gov/biblio/2283673},
journal = {Physical Review. D.},
issn = {ISSN 2470-0010},
number = {1},
volume = {109},
place = {United States},
publisher = {American Physical Society (APS)},
year = {2024},
month = {01}}
Univ. of Colorado, Boulder, CO (United States); Lawrence Berkeley National Laboratory (LBNL), Berkeley, CA (United States). National Energy Research Scientific Computing Center (NERSC); Oak Ridge National Laboratory (ORNL), Oak Ridge, TN (United States). Oak Ridge Leadership Computing Facility (OLCF)
Sponsoring Organization:
USDOE Office of Science (SC), High Energy Physics (HEP)
Grant/Contract Number:
SC0010005; AC02-05CH11231; AC05-00OR22725
OSTI ID:
2283673
Alternate ID(s):
OSTI ID: 2294139
Journal Information:
Physical Review. D., Journal Name: Physical Review. D. Journal Issue: 1 Vol. 109; ISSN 2470-0010
Proceedings of the Royal Society of London. Series A: Mathematical, Physical and Engineering Sciences, Vol. 453, Issue 1966https://doi.org/10.1098/rspa.1997.0122
Proceedings of the Royal Society of London. Series A: Mathematical, Physical and Engineering Sciences, Vol. 455, Issue 1985https://doi.org/10.1098/rspa.1999.0387