Estimation of partial least squares regression prediction uncertainty when the reference values carry a sizeable measurement error

J.A. Fernandez Pierna; L. Lin; F. Wahl; N.M. Faber; D.L. Massart

doi:10.1016/S0169-7439(02)00139-9

Estimation of partial least squares regression prediction uncertainty when the reference values carry a sizeable measurement error

J.A. Fernandez Pierna, L. Lin, F. Wahl, N.M. Faber, D.L. Massart

Research output: Contribution to journal › Article › Academic › peer-review

68 Citations (Scopus)

Abstract

The prediction uncertainty is studied when using a multivariate partial least squares regression (PLSR) model constructed with reference values that contain a sizeable measurement error. Several approximate expressions for calculating a sample-specific standard error of prediction have been proposed in the literature. In addition, Monte Carlo simulation methods such as the bootstrap and the noise addition method can give an estimate of this uncertainty. In this paper, two approximate expressions are compared with the simulation methods for three near-infrared data sets.

Original language	English
Pages (from-to)	281-291
Journal	Chemometrics and Intelligent Laboratory Systems
Volume	65
Issue number	2
DOIs	https://doi.org/10.1016/S0169-7439(02)00139-9
Publication status	Published - 2003

Keywords

principal component regression
multivariate calibration
confidence-intervals
models
unscrambler
construction
propagation
connection
validation
critique

Access to Document

10.1016/S0169-7439(02)00139-9

Cite this

@article{2ebb580015b94e5abffe139d14bbfff0,

title = "Estimation of partial least squares regression prediction uncertainty when the reference values carry a sizeable measurement error",

abstract = "The prediction uncertainty is studied when using a multivariate partial least squares regression (PLSR) model constructed with reference values that contain a sizeable measurement error. Several approximate expressions for calculating a sample-specific standard error of prediction have been proposed in the literature. In addition, Monte Carlo simulation methods such as the bootstrap and the noise addition method can give an estimate of this uncertainty. In this paper, two approximate expressions are compared with the simulation methods for three near-infrared data sets.",

keywords = "principal component regression, multivariate calibration, confidence-intervals, models, unscrambler, construction, propagation, connection, validation, critique",

author = "{Fernandez Pierna}, J.A. and L. Lin and F. Wahl and N.M. Faber and D.L. Massart",

year = "2003",

doi = "10.1016/S0169-7439(02)00139-9",

language = "English",

volume = "65",

pages = "281--291",

journal = "Chemometrics and Intelligent Laboratory Systems",

issn = "0169-7439",

publisher = "Elsevier",

number = "2",

}

TY - JOUR

T1 - Estimation of partial least squares regression prediction uncertainty when the reference values carry a sizeable measurement error

AU - Fernandez Pierna, J.A.

AU - Lin, L.

AU - Wahl, F.

AU - Faber, N.M.

AU - Massart, D.L.

PY - 2003

Y1 - 2003

N2 - The prediction uncertainty is studied when using a multivariate partial least squares regression (PLSR) model constructed with reference values that contain a sizeable measurement error. Several approximate expressions for calculating a sample-specific standard error of prediction have been proposed in the literature. In addition, Monte Carlo simulation methods such as the bootstrap and the noise addition method can give an estimate of this uncertainty. In this paper, two approximate expressions are compared with the simulation methods for three near-infrared data sets.

AB - The prediction uncertainty is studied when using a multivariate partial least squares regression (PLSR) model constructed with reference values that contain a sizeable measurement error. Several approximate expressions for calculating a sample-specific standard error of prediction have been proposed in the literature. In addition, Monte Carlo simulation methods such as the bootstrap and the noise addition method can give an estimate of this uncertainty. In this paper, two approximate expressions are compared with the simulation methods for three near-infrared data sets.

KW - principal component regression

KW - multivariate calibration

KW - confidence-intervals

KW - models

KW - unscrambler

KW - construction

KW - propagation

KW - connection

KW - validation

KW - critique

U2 - 10.1016/S0169-7439(02)00139-9

DO - 10.1016/S0169-7439(02)00139-9

M3 - Article

VL - 65

SP - 281

EP - 291

JO - Chemometrics and Intelligent Laboratory Systems

JF - Chemometrics and Intelligent Laboratory Systems

SN - 0169-7439

IS - 2

ER -

Estimation of partial least squares regression prediction uncertainty when the reference values carry a sizeable measurement error

Abstract

Keywords

Access to Document

Fingerprint

Cite this