Sequential Designs Based on Bayesian Uncertainty Quantification in Sparse Representation Surrogate Modeling

Ray Bing Chen, Weichung Wang, C. F.Jeff Wu

Research output: Contribution to journalArticle

2 Citations (Scopus)

Abstract

A numerical method, called overcomplete basis surrogate method (OBSM), was recently proposed, which employs overcomplete basis functions to achieve sparse representations. While the method can handle nonstationary response without the need of inverting large covariance matrices, it lacks the capability to quantify uncertainty in predictions. We address this issue by proposing a Bayesian approach that first imposes a normal prior on the large space of linear coefficients, then applies the Markov chain Monte Carlo (MCMC) algorithm to generate posterior samples for predictions. From these samples, Bayesian credible intervals can then be obtained to assess prediction uncertainty. A key application for the proposed method is the efficient construction of sequential designs. Several sequential design procedures with different infill criteria are proposed based on the generated posterior samples. Numerical studies show that the proposed schemes are capable of solving problems of positive point identification, optimization, and surrogate fitting.

Original languageEnglish
Pages (from-to)139-152
Number of pages14
JournalTechnometrics
Volume59
Issue number2
DOIs
Publication statusPublished - 2017 Apr 3

Fingerprint

Sequential Design
Uncertainty Quantification
Sparse Representation
Prediction
Modeling
Credible Interval
Uncertainty
Markov Chain Monte Carlo Algorithms
Covariance matrix
Bayesian Approach
Markov processes
Basis Functions
Numerical Study
Numerical methods
Quantify
Numerical Methods
Optimization
Coefficient

All Science Journal Classification (ASJC) codes

  • Statistics and Probability
  • Modelling and Simulation
  • Applied Mathematics

Cite this

@article{063baeb254694a8c90e1be2e423b460c,
title = "Sequential Designs Based on Bayesian Uncertainty Quantification in Sparse Representation Surrogate Modeling",
abstract = "A numerical method, called overcomplete basis surrogate method (OBSM), was recently proposed, which employs overcomplete basis functions to achieve sparse representations. While the method can handle nonstationary response without the need of inverting large covariance matrices, it lacks the capability to quantify uncertainty in predictions. We address this issue by proposing a Bayesian approach that first imposes a normal prior on the large space of linear coefficients, then applies the Markov chain Monte Carlo (MCMC) algorithm to generate posterior samples for predictions. From these samples, Bayesian credible intervals can then be obtained to assess prediction uncertainty. A key application for the proposed method is the efficient construction of sequential designs. Several sequential design procedures with different infill criteria are proposed based on the generated posterior samples. Numerical studies show that the proposed schemes are capable of solving problems of positive point identification, optimization, and surrogate fitting.",
author = "Chen, {Ray Bing} and Weichung Wang and Wu, {C. F.Jeff}",
year = "2017",
month = "4",
day = "3",
doi = "10.1080/00401706.2016.1172027",
language = "English",
volume = "59",
pages = "139--152",
journal = "Technometrics",
issn = "0040-1706",
publisher = "American Statistical Association",
number = "2",

}

Sequential Designs Based on Bayesian Uncertainty Quantification in Sparse Representation Surrogate Modeling. / Chen, Ray Bing; Wang, Weichung; Wu, C. F.Jeff.

In: Technometrics, Vol. 59, No. 2, 03.04.2017, p. 139-152.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Sequential Designs Based on Bayesian Uncertainty Quantification in Sparse Representation Surrogate Modeling

AU - Chen, Ray Bing

AU - Wang, Weichung

AU - Wu, C. F.Jeff

PY - 2017/4/3

Y1 - 2017/4/3

N2 - A numerical method, called overcomplete basis surrogate method (OBSM), was recently proposed, which employs overcomplete basis functions to achieve sparse representations. While the method can handle nonstationary response without the need of inverting large covariance matrices, it lacks the capability to quantify uncertainty in predictions. We address this issue by proposing a Bayesian approach that first imposes a normal prior on the large space of linear coefficients, then applies the Markov chain Monte Carlo (MCMC) algorithm to generate posterior samples for predictions. From these samples, Bayesian credible intervals can then be obtained to assess prediction uncertainty. A key application for the proposed method is the efficient construction of sequential designs. Several sequential design procedures with different infill criteria are proposed based on the generated posterior samples. Numerical studies show that the proposed schemes are capable of solving problems of positive point identification, optimization, and surrogate fitting.

AB - A numerical method, called overcomplete basis surrogate method (OBSM), was recently proposed, which employs overcomplete basis functions to achieve sparse representations. While the method can handle nonstationary response without the need of inverting large covariance matrices, it lacks the capability to quantify uncertainty in predictions. We address this issue by proposing a Bayesian approach that first imposes a normal prior on the large space of linear coefficients, then applies the Markov chain Monte Carlo (MCMC) algorithm to generate posterior samples for predictions. From these samples, Bayesian credible intervals can then be obtained to assess prediction uncertainty. A key application for the proposed method is the efficient construction of sequential designs. Several sequential design procedures with different infill criteria are proposed based on the generated posterior samples. Numerical studies show that the proposed schemes are capable of solving problems of positive point identification, optimization, and surrogate fitting.

UR - http://www.scopus.com/inward/record.url?scp=85017549525&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85017549525&partnerID=8YFLogxK

U2 - 10.1080/00401706.2016.1172027

DO - 10.1080/00401706.2016.1172027

M3 - Article

AN - SCOPUS:85017549525

VL - 59

SP - 139

EP - 152

JO - Technometrics

JF - Technometrics

SN - 0040-1706

IS - 2

ER -