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

Title: On the Bayesian treed multivariate Gaussian process with linear model of coregionalization

; ;
Publication Date:
Sponsoring Org.:
OSTI Identifier:
Resource Type:
Journal Article: Publisher's Accepted Manuscript
Journal Name:
Journal of Statistical Planning and Inference
Additional Journal Information:
Journal Volume: 157-158; Journal Issue: C; Related Information: CHORUS Timestamp: 2016-09-04 21:12:27; Journal ID: ISSN 0378-3758
Country of Publication:
Country unknown/Code not available

Citation Formats

Konomi, Bledar, Karagiannis, Georgios, and Lin, Guang. On the Bayesian treed multivariate Gaussian process with linear model of coregionalization. Country unknown/Code not available: N. p., 2015. Web. doi:10.1016/j.jspi.2014.08.010.
Konomi, Bledar, Karagiannis, Georgios, & Lin, Guang. On the Bayesian treed multivariate Gaussian process with linear model of coregionalization. Country unknown/Code not available. doi:10.1016/j.jspi.2014.08.010.
Konomi, Bledar, Karagiannis, Georgios, and Lin, Guang. 2015. "On the Bayesian treed multivariate Gaussian process with linear model of coregionalization". Country unknown/Code not available. doi:10.1016/j.jspi.2014.08.010.
title = {On the Bayesian treed multivariate Gaussian process with linear model of coregionalization},
author = {Konomi, Bledar and Karagiannis, Georgios and Lin, Guang},
abstractNote = {},
doi = {10.1016/j.jspi.2014.08.010},
journal = {Journal of Statistical Planning and Inference},
number = C,
volume = 157-158,
place = {Country unknown/Code not available},
year = 2015,
month = 2

Journal Article:
Free Publicly Available Full Text
Publisher's Version of Record at 10.1016/j.jspi.2014.08.010

Citation Metrics:
Cited by: 1work
Citation information provided by
Web of Science

Save / Share:
  • The Bayesian treed Gaussian process (BTGP) has gained popularity in recent years because it provides a straightforward mechanism for modeling non-stationary data and can alleviate computational demands by fitting models to less data. The extension of BTGP to the multivariate setting requires us to model the cross-covariance and to propose efficient algorithms that can deal with trans-dimensional MCMC moves. In this paper we extend the cross-covariance of the Bayesian treed multivariate Gaussian process (BTMGP) to that of linear model of Coregionalization (LMC) cross-covariances. Different strategies have been developed to improve the MCMC mixing and invert smaller matrices in the Bayesianmore » inference. Moreover, we compare the proposed BTMGP with existing multiple BTGP and BTMGP in test cases and multiphase flow computer experiment in a full scale regenerator of a carbon capture unit. The use of the BTMGP with LMC cross-covariance helped to predict the computer experiments relatively better than existing competitors. The proposed model has a wide variety of applications, such as computer experiments and environmental data. In the case of computer experiments we also develop an adaptive sampling strategy for the BTMGP with LMC cross-covariance function.« less
  • Computer experiments (numerical simulations) are widely used in scientific research to study and predict the behavior of complex systems, which usually have responses consisting of a set of distinct outputs. The computational cost of the simulations at high resolution are often expensive and become impractical for parametric studies at different input values. To overcome these difficulties we develop a Bayesian treed multivariate Gaussian process (BTMGP) as an extension of the Bayesian treed Gaussian process (BTGP) in order to model and evaluate a multivariate process. A suitable choice of covariance function and the prior distributions facilitates the different Markov chain Montemore » Carlo (MCMC) movements. We utilize this model to sequentially sample the input space for the most informative values, taking into account model uncertainty and expertise gained. A simulation study demonstrates the use of the proposed method and compares it with alternative approaches. We apply the sequential sampling technique and BTMGP to model the multiphase flow in a full scale regenerator of a carbon capture unit. The application presented in this paper is an important tool for research into carbon dioxide emissions from thermal power plants.« less
  • This paper extends the results of Masreliez on the design of non-Gaussian estimators for a more general class of the parameter estimation problem when the system state and the observation noise may be dependent and non-Gaussian simultaneously. It is shown that the proposed non-Gaussian algorithms can approximate with high precision the minimum mean square estimator. Application of the approach to the design of different optimal (and stable) estimation algorithms is illustrated. The efficiency of the proposed algorithms is tested in some simulation experiments.
  • Surrogate models are commonly used in Bayesian approaches such as Markov Chain Monte Carlo (MCMC) to avoid repetitive CPU-demanding model evaluations. However, the approximation error of a surrogate may lead to biased estimations of the posterior distribution. This bias can be corrected by constructing a very accurate surrogate or implementing MCMC in a two-stage manner. Since the two-stage MCMC requires extra original model evaluations, the computational cost is still high. If the information of measurement is incorporated, a locally accurate approximation of the original model can be adaptively constructed with low computational cost. Based on this idea, we propose amore » Gaussian process (GP) surrogate-based Bayesian experimental design and parameter estimation approach for groundwater contaminant source identification problems. A major advantage of the GP surrogate is that it provides a convenient estimation of the approximation error, which can be incorporated in the Bayesian formula to avoid over-confident estimation of the posterior distribution. The proposed approach is tested with a numerical case study. Without sacrificing the estimation accuracy, the new approach achieves about 200 times of speed-up compared to our previous work using two-stage MCMC.« less
  • In cases where field or experimental measurements are not available, computer models can model real physical or engineering systems to reproduce their outcomes. They are usually calibrated in light of experimental data to create a better representation of the real system. Statistical methods, based on Gaussian processes, for calibration and prediction have been especially important when the computer models are expensive and experimental data limited. In this paper, we develop the Bayesian treed calibration (BTC) as an extension of standard Gaussian process calibration methods to deal with non-stationarity computer models and/or their discrepancy from the field (or experimental) data. Ourmore » proposed method partitions both the calibration and observable input space, based on a binary tree partitioning, into sub-regions where existing model calibration methods can be applied to connect a computer model with the real system. The estimation of the parameters in the proposed model is carried out using Markov chain Monte Carlo (MCMC) computational techniques. Different strategies have been applied to improve mixing. We illustrate our method in two artificial examples and a real application that concerns the capture of carbon dioxide with AX amine based sorbents. The source code and the examples analyzed in this paper are available as part of the supplementary materials.« less