Maximum likelihood estimation for Gaussian processes under inequality constraints
From MaRDI portal
Publication:2323946
DOI10.1214/19-EJS1587zbMATH Open1428.62420arXiv1804.03378OpenAlexW2972250097MaRDI QIDQ2323946FDOQ2323946
Authors: François Bachoc, Agnès Lagnoux, Andrés F. López-Lopera 
Publication date: 13 September 2019
Published in: Electronic Journal of Statistics (Search for Journal in Brave)
Abstract: We consider covariance parameter estimation for a Gaussian process under inequality constraints (boundedness, monotonicity or convexity) in fixed-domain asymptotics. We address the estimation of the variance parameter and the estimation of the microergodic parameter of the Mat'ern and Wendland covariance functions. First, we show that the (unconstrained) maximum likelihood estimator has the same asymptotic distribution, unconditionally and conditionally to the fact that the Gaussian process satisfies the inequality constraints. Then, we study the recently suggested constrained maximum likelihood estimator. We show that it has the same asymptotic distribution as the (unconstrained) maximum likelihood estimator. In addition, we show in simulations that the constrained maximum likelihood estimator is generally more accurate on finite samples. Finally, we provide extensions to prediction and to noisy observations.
Full work available at URL: https://arxiv.org/abs/1804.03378
Recommendations
- Maximum likelihood estimation for a bivariate Gaussian process under fixed domain asymptotics
- Composite likelihood estimation for a Gaussian process under fixed domain asymptotics
- Asymptotic properties of a maximum likelihood estimator with data from a Gaussian process
- Gaussian processes with linear operator inequality constraints
- Finite-Dimensional Gaussian Approximation with Linear Inequality Constraints
asymptotic normalityGaussian processesinequality constraintsfixed-domain asymptoticsconstrained maximum likelihood
Asymptotic properties of parametric estimators (62F12) Inference from spatial processes (62M30) Gaussian processes (60G15)
Cites Work
- Efficient global optimization of expensive black-box functions
- Asymptotic Statistics
- Fixed-domain asymptotic properties of tapered maximum likelihood estimators
- Asymptotically efficient prediction of a random field with a missspecified covariance function
- Interpolation of spatial data. Some theory for kriging
- The design and analysis of computer experiments.
- Design and analysis of computer experiments. With comments and a rejoinder by the authors
- Estimation and prediction using generalized Wendland covariance functions under fixed domain asymptotics
- Compactly supported correlation functions
- On fixed-domain asymptotics and covariance tapering in Gaussian random field models
- Gaussian processes for machine learning.
- The role of the range parameter for estimation and prediction in geostatistics
- Inconsistent Estimation and Asymptotically Equal Interpolations in Model-Based Geostatistics
- Title not available (Why is that?)
- Uniform asymptotic optimality of linear predictions of a random field using an incorrect second-order structure
- Quadratic variations and estimation of the local Hölder index of a Gaussian process
- Calibration of computer models with multivariate output
- A supermartingale approach to Gaussian process based sequential design of experiments
- Approximations for binary Gaussian process classification
- Level Sets and Extrema of Random Processes and Fields
- Title not available (Why is that?)
- Random Fields and Geometry
- On the consistent separation of scale and variance for Gaussian random fields
- A comparison of generalized cross validation and modified maximum likelihood for estimating the parameters of a stochastic process
- Title not available (Why is that?)
- Asymptotic analysis of the role of spatial sampling for covariance parameter estimation of Gaussian processes
- Gaussian process emulators for computer experiments with inequality constraints
- Kriging of financial term-structures
- Cross validation and maximum likelihood estimations of hyper-parameters of Gaussian processes with model misspecification
- Monotone Emulation of Computer Experiments
- Asymptotic properties of a maximum likelihood estimator with data from a Gaussian process
- Title not available (Why is that?)
- On the effect of covariance function estimation on the accuracy of kriging predictors
- Maximum likelihood estimation of parameters under a spatial sampling scheme
- Bounds on the efficiency of linear predictions using an incorrect covariance function
- Title not available (Why is that?)
- Scattered Data Approximation
- Mixed domain asymptotics for a stochastic process model with time trend and measurement error
- Derivative reproducing properties for kernel methods in learning theory
- Asymptotic analysis of covariance parameter estimation for Gaussian processes in the misspecified case
- Gaussian process modeling with inequality constraints
- Finite-Dimensional Gaussian Approximation with Linear Inequality Constraints
Cited In (12)
- A matrix-free approach for solving the parametric Gaussian process maximum likelihood problem
- Maximum Likelihood Estimation and Uncertainty Quantification for Gaussian Process Approximation of Deterministic Functions
- Short Communication: Beyond Surrogate Modeling: Learning the Local Volatility via Shape Constraints
- Asymptotic Analysis of Maximum Likelihood Estimation of Covariance Parameters for Gaussian Processes: An Introduction with Proofs
- Finite-dimensional approximation of Gaussian processes with linear inequality constraints and noisy observations
- Exact maximum likelihood for incomplete data from a correlated gaussian process
- Asymptotic behavior of the likelihood function of covariance matrices of spatial Gaussian processes
- Maximum likelihood estimation in Gaussian models under total positivity
- Asymptotic properties of a maximum likelihood estimator with data from a Gaussian process
- Gaussian field on the symmetric group: prediction and learning
- Composite likelihood estimation for a Gaussian process under fixed domain asymptotics
- Sequential Construction and Dimension Reduction of Gaussian Processes Under Inequality Constraints
Uses Software
This page was built for publication: Maximum likelihood estimation for Gaussian processes under inequality constraints
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2323946)
