skip to main content
OSTI.GOV title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: Evaluating marginal likelihood with thermodynamic integration method and comparison with several other numerical methods

Journal Article · · Water Resources Research
DOI:https://doi.org/10.1002/2014WR016718· OSTI ID:1326496
 [1];  [2];  [2];  [2];  [3];  [4];  [5]
  1. Hefei Univ. of Technology, Hefei (China); Florida State Univ., Tallahassee, FL (United States)
  2. Florida State Univ., Tallahassee, FL (United States)
  3. Nanjing Univ., Nanjing (China)
  4. Oak Ridge National Lab. (ORNL), Oak Ridge, TN (United States)
  5. Hefei Univ. of Technology, Hefei (China)
+ Show Author Affiliations

Abstract Evaluating marginal likelihood is the most critical and computationally expensive task, when conducting Bayesian model averaging to quantify parametric and model uncertainties. The evaluation is commonly done by using Laplace approximations to evaluate semianalytical expressions of the marginal likelihood or by using Monte Carlo (MC) methods to evaluate arithmetic or harmonic mean of a joint likelihood function. This study introduces a new MC method, i.e., thermodynamic integration, which has not been attempted in environmental modeling. Instead of using samples only from prior parameter space (as in arithmetic mean evaluation) or posterior parameter space (as in harmonic mean evaluation), the thermodynamic integration method uses samples generated gradually from the prior to posterior parameter space. This is done through a path sampling that conducts Markov chain Monte Carlo simulation with different power coefficient values applied to the joint likelihood function. The thermodynamic integration method is evaluated using three analytical functions by comparing the method with two variants of the Laplace approximation method and three MC methods, including the nested sampling method that is recently introduced into environmental modeling. The thermodynamic integration method outperforms the other methods in terms of their accuracy, convergence, and consistency. The thermodynamic integration method is also applied to a synthetic case of groundwater modeling with four alternative models. The application shows that model probabilities obtained using the thermodynamic integration method improves predictive performance of Bayesian model averaging. The thermodynamic integration method is mathematically rigorous, and its MC implementation is computationally general for a wide range of environmental problems.

Research Organization:
Oak Ridge National Laboratory (ORNL), Oak Ridge, TN (United States)
Sponsoring Organization:
USDOE
Grant/Contract Number:
AC05-00OR22725; DE‐SC0008272
OSTI ID:
1326496
Alternate ID(s):
OSTI ID: 1402380
Journal Information:
Water Resources Research, Vol. 52, Issue 2; ISSN 0043-1397
Publisher:
American Geophysical Union (AGU)Copyright Statement
Country of Publication:
United States
Language:
English
Citation Metrics:
Cited by: 38 works
Citation information provided by
Web of Science

References (43)

Bayes Factors journal June 1995
Maximum likelihood Bayesian averaging of spatial variability models in unsaturated fractured tuff: MAXIMUM LIKELIHOOD BAYESIAN MODEL AVERAGING journal May 2004
Comment on “Inverse groundwater modeling for hydraulic conductivity estimation using Bayesian model averaging and variance window” by Frank T.-C. Tsai and Xiaobao Li: COMMENTARY journal February 2010
Nested Sampling
  • Skilling, John
  • BAYESIAN INFERENCE AND MAXIMUM ENTROPY METHODS IN SCIENCE AND ENGINEERING: 24th International Workshop on Bayesian Inference and Maximum Entropy Methods in Science and Engineering, AIP Conference Proceedings https://doi.org/10.1063/1.1835238
conference January 2004
Marginal likelihood estimation via power posteriors journal July 2008
Simulating normalizing constants: from importance sampling to bridge sampling to path sampling journal May 1998
Nested sampling algorithm for subsurface flow model selection, uncertainty quantification, and nonlinear calibration: NESTED SAMPLING FOR SUBSURFACE FLOW MODELS journal December 2013
Treatment of input uncertainty in hydrologic modeling: Doing hydrology backward with Markov chain Monte Carlo simulation: FORCING DATA ERROR USING MCMC SAMPLING journal December 2008
Conceptual model uncertainty in groundwater modeling: Combining generalized likelihood uncertainty estimation and Bayesian model averaging: EVALUATING CONCEPTUAL MODEL journal December 2008
Hydrological model selection: A Bayesian alternative: HYDROLOGICAL MODEL SELECTION journal October 2005
Review of surrogate modeling in water resources: REVIEW journal July 2012
Assessment of parametric uncertainty for groundwater reactive transport modeling journal May 2014
A multimodel data assimilation framework via the ensemble Kalman filter journal May 2014
Estimation of Aquifer Parameters Under Transient and Steady State Conditions: 1. Maximum Likelihood Method Incorporating Prior Information journal February 1986
Assessing five evolving microbial enzyme models against field measurements from a semiarid savannah-What are the mechanisms of soil respiration pulses? journal September 2014
Bayesian analysis of data-worth considering model and parameter uncertainties journal February 2012
A Model-Averaging Method for Assessing Groundwater Conceptual Model Uncertainty journal August 2010
Analysis of regression confidence intervals and Bayesian credible intervals for uncertainty quantification: REGRESSION CONFIDENCE AND BAYESIAN CREDIBLE INTERVALS journal September 2012
Computing the Bayes Factor from a Markov Chain Monte Carlo Simulation of the Posterior Distribution journal September 2012
Hierarchical Bayesian model averaging for hydrostratigraphic modeling: Uncertainty segregation and comparative evaluation: HIERARCHICAL BAYESIAN MODEL AVERAGING journal September 2013
An adaptive sparse-grid high-order stochastic collocation method for Bayesian inference in groundwater reactive transport modeling: Sparse-Grid Method for Bayesian Inference journal October 2013
Efficient posterior exploration of a high-dimensional groundwater model from two-stage Markov chain Monte Carlo simulation and polynomial chaos expansion: Speeding up MCMC Simulation of a Groundwater Model journal May 2013
Estimating the evidence - a review journal January 2012
Model selection on solid ground: Rigorous comparison of nine ways to evaluate Bayesian model evidence journal December 2014
Postaudit evaluation of conceptual model uncertainty for a glacial aquifer groundwater flow and contaminant transport model journal January 2010
High-dimensional posterior exploration of hydrologic models using multiple-try DREAM (ZS) and high-performance computing : EFFICIENT MCMC FOR HIGH-DIMENSIONAL PROBLEMS journal January 2012
Unified Framework to Evaluate Panmixia and Migration Direction Among Multiple Sampling Locations journal February 2010
Computing Bayes Factors Using Thermodynamic Integration journal April 2006
Hydrologic data assimilation using particle Markov chain Monte Carlo simulation: Theory, concepts and applications journal January 2013
Maximum likelihood Bayesian averaging of uncertain model predictions journal November 2003
On model selection criteria in multimodel analysis: ON MODEL SELECTION CRITERIA IN MULTIMODEL ANALYSIS journal March 2008
A Bayesian approach for inverse modeling, data assimilation, and conditional simulation of spatial random fields journal October 2010
Constructive epistemic modeling of groundwater flow with geological structure and boundary condition uncertainty under the Bayesian paradigm journal September 2014
Marginal Likelihood From the Metropolis–Hastings Output journal March 2001
Improving Marginal Likelihood Estimation for Bayesian Phylogenetic Model Selection journal December 2010
Markov Chain Monte Carlo Methods for Computing Bayes Factors: A Comparative Review journal September 2001
Multimodel Bayesian analysis of data-worth applied to unsaturated fractured tuffs journal January 2012
Model complexity control for hydrologic prediction: MODEL COMPLEXITY CONTROL journal August 2008
Using MCMC chain outputs to efficiently estimate Bayes factors journal October 2011
Dependence of Bayesian Model Selection Criteria and Fisher Information Matrix on Sample Size journal October 2011
Effects of error covariance structure on estimation of model averaging weights and predictive performance: EFFECTS OF ERROR COVARIANCE STRUCTURE ON MODEL AVERAGING journal September 2013
David Draper and E. I. George, and a rejoinder by the authors journal November 1999
Markov Chain Sampling Methods for Dirichlet Process Mixture Models journal June 2000

Cited By (2)

Hydrogeological Model Selection Among Complex Spatial Priors journal August 2019
Relative model score: a scoring rule for evaluating ensemble simulations with application to microbial soil respiration modeling journal August 2018

Similar Records

Related Subjects

97 MATHEMATICS AND COMPUTING
Bayesian model averaging
arithmetic mean
harmonic mean
nested sampling
model uncertainty
Markov chain Monte Carlo