Gaussian process regression with functional covariates and multivariate response

B. Wang, T. Chen, Aiping Xu

Research output: Contribution to journalArticle

31 Downloads (Pure)

Abstract

Gaussian process regression (GPR) has been shown to be a powerful and effective nonparametric method for regression, classification and interpolation, due to many of its desirable properties. However, most GPR models consider univariate or multivariate covariates only. In this paper we extend the GPR models to cases where the covariates include both functional and multivariate variables and the response is multidimensional. The model naturally incorporates two different types of covariates: multivariate and functional, and the principal component analysis is used to de-correlate the multivariate response which avoids the widely recognised difficulty in the multi-output GPR models of formulating covariance functions which have to describe the correlations not only between data points but also between responses. The usefulness of the proposed method is demonstrated through a simulated example and two real data sets in chemometrics.
Original languageEnglish
Pages (from-to)1-6
Number of pages5
JournalChemometrics and Intelligent Laboratory Systems
Volume163
DOIs
Publication statusPublished - 3 Feb 2017

Fingerprint

Principal component analysis
Interpolation

Bibliographical note

Due to publisher policy the full text is not available on the repository until the 3rd of February 2018

Keywords

  • Gaussian process regression
  • Functional data analysis
  • Functional covariates
  • Multivariate response
  • Semi-metrics

Cite this

Gaussian process regression with functional covariates and multivariate response. / Wang, B.; Chen, T.; Xu, Aiping.

In: Chemometrics and Intelligent Laboratory Systems, Vol. 163, 03.02.2017, p. 1-6.

Research output: Contribution to journalArticle

@article{54087bfa0d734c99bd730bda41f1408f,
title = "Gaussian process regression with functional covariates and multivariate response",
abstract = "Gaussian process regression (GPR) has been shown to be a powerful and effective nonparametric method for regression, classification and interpolation, due to many of its desirable properties. However, most GPR models consider univariate or multivariate covariates only. In this paper we extend the GPR models to cases where the covariates include both functional and multivariate variables and the response is multidimensional. The model naturally incorporates two different types of covariates: multivariate and functional, and the principal component analysis is used to de-correlate the multivariate response which avoids the widely recognised difficulty in the multi-output GPR models of formulating covariance functions which have to describe the correlations not only between data points but also between responses. The usefulness of the proposed method is demonstrated through a simulated example and two real data sets in chemometrics.",
keywords = "Gaussian process regression, Functional data analysis, Functional covariates, Multivariate response, Semi-metrics",
author = "B. Wang and T. Chen and Aiping Xu",
note = "Due to publisher policy the full text is not available on the repository until the 3rd of February 2018",
year = "2017",
month = "2",
day = "3",
doi = "10.1016/j.chemolab.2017.02.001",
language = "English",
volume = "163",
pages = "1--6",
journal = "Chemometrics and Intelligent Laboratory Systems",
issn = "0169-7439",
publisher = "Elsevier",

}

TY - JOUR

T1 - Gaussian process regression with functional covariates and multivariate response

AU - Wang, B.

AU - Chen, T.

AU - Xu, Aiping

N1 - Due to publisher policy the full text is not available on the repository until the 3rd of February 2018

PY - 2017/2/3

Y1 - 2017/2/3

N2 - Gaussian process regression (GPR) has been shown to be a powerful and effective nonparametric method for regression, classification and interpolation, due to many of its desirable properties. However, most GPR models consider univariate or multivariate covariates only. In this paper we extend the GPR models to cases where the covariates include both functional and multivariate variables and the response is multidimensional. The model naturally incorporates two different types of covariates: multivariate and functional, and the principal component analysis is used to de-correlate the multivariate response which avoids the widely recognised difficulty in the multi-output GPR models of formulating covariance functions which have to describe the correlations not only between data points but also between responses. The usefulness of the proposed method is demonstrated through a simulated example and two real data sets in chemometrics.

AB - Gaussian process regression (GPR) has been shown to be a powerful and effective nonparametric method for regression, classification and interpolation, due to many of its desirable properties. However, most GPR models consider univariate or multivariate covariates only. In this paper we extend the GPR models to cases where the covariates include both functional and multivariate variables and the response is multidimensional. The model naturally incorporates two different types of covariates: multivariate and functional, and the principal component analysis is used to de-correlate the multivariate response which avoids the widely recognised difficulty in the multi-output GPR models of formulating covariance functions which have to describe the correlations not only between data points but also between responses. The usefulness of the proposed method is demonstrated through a simulated example and two real data sets in chemometrics.

KW - Gaussian process regression

KW - Functional data analysis

KW - Functional covariates

KW - Multivariate response

KW - Semi-metrics

U2 - 10.1016/j.chemolab.2017.02.001

DO - 10.1016/j.chemolab.2017.02.001

M3 - Article

VL - 163

SP - 1

EP - 6

JO - Chemometrics and Intelligent Laboratory Systems

JF - Chemometrics and Intelligent Laboratory Systems

SN - 0169-7439

ER -