Nested kriging predictions for datasets with a large number of observations
From MaRDI portal
Publication:1704020
Abstract: This work falls within the context of predicting the value of a real function at some input locations given a limited number of observations of this function. The Kriging interpolation technique (or Gaussian process regression) is often considered to tackle such a problem but the method suffers from its computational burden when the number of observation points is large. We introduce in this article nested Kriging predictors which are constructed by aggregating sub-models based on subsets of observation points. This approach is proven to have better theoretical properties than other aggregation methods that can be found in the literature. Contrarily to some other methods it can be shown that the proposed aggregation method is consistent. Finally, the practical interest of the proposed method is illustrated on simulated datasets and on an industrial test case with observations in a 6-dimensional space.
Recommendations
- Kriging with large data sets using sparse matrix techniques
- Subset selection from large datasets for kriging modeling
- Kriging prediction for spatial random fields valued in a Hilbert space
- An empirical comparison of kriging methods for nonlinear spatial point prediction
- Fixed Rank Kriging for Very Large Spatial Data Sets
- Kriging analysis for classification of spatial and multivariate observations
- Spatial prediction and ordinary kriging
Cites work
- A Fast, Optimal Spatial-Prediction Method for Massive Datasets
- A well-conditioned and sparse estimation of covariance and inverse covariance matrices using a joint penalty
- Bayesian nonstationary spatial modeling for very large datasets
- Combining Probability Distributions from Dependent Information Sources
- Combining probability distributions: A critique and an annotated bibliography
- Combining probability forecasts
- Cross validation and maximum likelihood estimations of hyper-parameters of Gaussian processes with model misspecification
- Full-scale approximations of spatio-temporal covariance models for large datasets
- Gaussian Markov Random Fields
- Gaussian processes for machine learning.
- scientific article; zbMATH DE number 45848 (Why is no real title available?)
- scientific article; zbMATH DE number 6159604 (Why is no real title available?)
- Mechanics of Solid Materials
- Properties and comparison of some kriging sub-model aggregation methods
- Stochastic recursive algorithms for optimization. Simultaneous perturbation methods
- String and membrane Gaussian processes
- The design and analysis of computer experiments.
- Training Products of Experts by Minimizing Contrastive Divergence
Cited in
(18)- A Bayesian optimization approach to find Nash equilibria
- Composite likelihood estimation for a Gaussian process under fixed domain asymptotics
- Analyzing stochastic computer models: a review with opportunities
- Nested aggregation of experts using inducing points for approximated Gaussian process regression
- Fixed-domain asymptotic properties of maximum composite likelihood estimators for Gaussian processes
- Properties and comparison of some kriging sub-model aggregation methods
- Learning with partition of unity-based kriging estimators
- Subset selection from large datasets for kriging modeling
- Gaussian processes for computer experiments
- Semi-parametric estimation of the variogram scale parameter of a Gaussian process with stationary increments
- Correlated product of experts for sparse Gaussian process regression
- Enhanced kriging leave-one-out cross-validation in improving model estimation and optimization
- Leveraging viscous Hamilton-Jacobi PDEs for uncertainty quantification in scientific machine learning
- Combination of optimization-free Kriging models for high-dimensional problems
- A model aggregation approach for high-dimensional large-scale optimization
- Incorporating subsampling into Bayesian models for high-dimensional spatial data
- A joint kriging model with application to constrained classification
- Hierarchical Bayesian Quadrature
This page was built for publication: Nested kriging predictions for datasets with a large number of observations
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1704020)
