Comparison of Gaussian process modeling software

DOI10.1016/J.EJOR.2017.10.002MaRDI QIDQ65764zbMATH OpenOpenAlexFDO

Authors Collin B. Erickson, Bruce E. Ankenman, Susan M. Sanchez, Collin B. Erickson, Bruce E. Ankenman, Susan M. Sanchez

Publication date April 2018

Published in European Journal of Operational Research (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/1710.03157

zbMATH Keywords

computer experimentsGaussian processesmetamodelssimulationstochastic kriging

Mathematics Subject Classification ID

Applications of statistics (62Pxx) Gaussian processes (60G15) Software, source code, etc. for problems pertaining to statistics (62-04) Design of statistical experiments (62Kxx)

Abstract: Gaussian process fitting, or kriging, is often used to create a model from a set of data. Many available software packages do this, but we show that very different results can be obtained from different packages even when using the same data and model. We describe the parameterization, features, and optimization used by eight different fitting packages that run on four different platforms. We then compare these eight packages using various data functions and data sets, revealing that there are stark differences between the packages. In addition to comparing the prediction accuracy, the predictive variance--which is important for evaluating precision of predictions and is often used in stopping criteria--is also evaluated.

Recommendations

• Kriging: methods and applications
• GPLP: a local and parallel computation toolbox for Gaussian process regression
• Gaussian processes for machine learning (GPML) toolbox
• Analysis methods for computer experiments: how to assess and what counts?
• GPstuff: Bayesian modeling with Gaussian processes

Cites work

• A Limited Memory Algorithm for Bound Constrained Optimization
• Analysis methods for computer experiments: how to assess and what counts?
• Bayesian Design and Analysis of Computer Experiments: Use of Derivatives in Surface Prediction
• Bayesian treed Gaussian process models with an application to computer modeling
• Cases for the nugget in modeling computer experiments
• Design and analysis of computer experiments. With comments and a rejoinder by the authors
• Design and analysis of simulation experiments
• Efficient global optimization of expensive black-box functions
• Gaussian processes for machine learning.
• On the limited memory BFGS method for large scale optimization
• Scikit-learn: machine learning in Python
• Stochastic kriging for simulation metamodeling
• The design and analysis of computer experiments.

Cited in

(9)

• On surrogate learning for linear stability assessment of Navier-Stokes equations with stochastic viscosity.
• Modeling and Active Learning for Experiments with Quantitative-Sequence Factors
• Comparing emulation methods for a high-resolution storm surge model
• Statistical Tests for Cross-Validation of Kriging Models
• An extended two-stage sequential optimization approach: properties and performance
• Practical Heteroscedastic Gaussian Process Modeling for Large Simulation Experiments
• GauPro
• Kriging: methods and applications
• Faster Kriging: facing high-dimensional simulators

Describes a project that uses

Uses Software

• JMP
• R
• L-BFGS
• DACE
• LBFGS-B
• SciPy
• rgenoud
• laGP
• EGO
• DiceKriging
• DiceOptim
• tgp
• Scikit
• mlegp
• GPfit
• GPy
• MaxPro
• Python
• Virtual library

This page was built for publication: Comparison of Gaussian process modeling software

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q65764)