Does Bayesian Model Averaging improve polynomial extrapolations? Two toy problems as tests

Abstract

We assess the accuracy of Bayesian polynomial extrapolations from small parameter values, x, to large values of x. We consider a set of polynomials of fixed order, intended as a proxy for a fixed-order effective field theory (EFT) description of data. We employ Bayesian Model Averaging (BMA) to combine results from different order polynomials (EFT orders). Our study considers two "toy problems" where the underlying function used to generate data sets is known. We use Bayesian parameter estimation to extract the polynomial coefficients that describe these data at low x. A "naturalness" prior is imposed on the coefficients, so that they are O(1). We Bayesian-Model-Average different polynomial degrees by weighting each according to its Bayesian evidence and compare the predictive performance of this Bayesian Model Average with that of the individual polynomials. The credibility intervals on the BMA forecast have the stated coverage properties more consistently than does the highest evidence polynomial, though BMA does not necessarily outperform every polynomial.

arXiv comment19 pages, 5 figures, v2 corresponds to published version

References in corpus (21)

• Using stacking to average Bayesian predictive distributions
• How well do we know the neutron-matter equation of state at the densities inside neutron stars? A Bayesian approach with correlated uncertainties
• Quantifying truncation errors in effective field theory
• Bayesian approach to model-based extrapolation of nuclear observables
• Quantifying Correlated Truncation Errors in Effective Field Theory
• Bayesian truncation errors in chiral effective field theory: nucleon-nucleon observables
• Neutron drip line in the Ca region from Bayesian model averaging
• Quantifying uncertainties and correlations in the nuclear-matter equation of state
• Bayesian parameter estimation for effective field theories
• Quantified limits of the nuclear landscape
• Get on the BAND Wagon: A Bayesian Framework for Quantifying Model Uncertainties in Nuclear Dynamics
• Exploring Bayesian parameter estimation for chiral effective field theory using nucleon-nucleon phase shifts
• Bayesian Methods for Parameter Estimation in Effective Field Theories
• Extraction of the neutron charge radius from a precision calculation of the deuteron structure radius
• High-accuracy calculation of the deuteron charge and quadrupole form factors in chiral effective field theory
• Precision determination of pion-nucleon coupling constants using effective field theory
• Probabilistic definition of the perturbative theoretical uncertainty from missing higher orders
• Statistical aspects of nuclear mass models
• Nucleon Polarisabilities at and Beyond Physical Pion Masses
• Bayesian analysis of capture reactions ${}^3\mathrm{He}(\alpha ,\gamma {)}^7\mathrm{Be}$ and ${}^3\mathrm{H}(\alpha ,\gamma {)}^7\mathrm{Li}$
• Holes in Bayesian Statistics

Cited by in corpus (3)

• Machine Learning in Nuclear Physics
• Improved information criteria for Bayesian model averaging in lattice field theory
• Gaussian process error modeling for chiral effective-field-theory calculations of $np\leftrightarrow d\gamma$ at low energies