Multivariate receptor models and model uncertainty

Eun Sug Park; Man Suk Oh; Peter Guttorp

doi:10.1016/S0169-7439(01)00185-X

Multivariate receptor models and model uncertainty

Eun Sug Park, Man Suk Oh, Peter Guttorp

Department of Statistics

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

45 Scopus citations

Abstract

Estimation of the number of major pollution sources, the source composition profiles, and the source contributions are the main interests in multivariate receptor modeling. Due to lack of identifiability of the receptor model, however, the estimation cannot be done without some additional assumptions. A common approach to this problem is to estimate the number of sources, q, at the first stage, and then estimate source profiles and contributions at the second stage, given additional constraints (identifiability conditions) to prevent source rotation/transformation and the assumption that the q-source model is correct. These assumptions on the parameters (the number of sources and identifiability conditions) are the main source of model uncertainty in multivariate receptor modeling. In this paper, we suggest a Bayesian approach to deal with model uncertainties in multivariate receptor models by using Markov chain Monte Carlo (MCMC) schemes. Specifically, we suggest a method which can simultaneously estimate parameters (compositions and contributions), parameter uncertainties, and model uncertainties (number of sources and identifiability conditions). Simulation results and an application to air pollution data are presented.

Original language	English
Pages (from-to)	49-67
Number of pages	19
Journal	Chemometrics and Intelligent Laboratory Systems
Volume	60
Issue number	1-2
DOIs	https://doi.org/10.1016/S0169-7439(01)00185-X
State	Published - 28 Jan 2002

Bibliographical note

Funding Information:
Although the research described in this article has been funded by the United States Environmental Protection Agency through agreement CR825173-01-0 to the University of Washington, it has not been subjected to the agency's required peer and policy review and therefore does not necessarily reflect the views of the Agency and no official endorsement should be inferred.

Keywords

Factor analysis models
Latent variable models
Marginal likelihood
Model identifiability
Model uncertainty
Number of sources
Posterior model probability

Access to Document

10.1016/S0169-7439(01)00185-X

Cite this

@article{c317b71bee43490b8cfc97cd672d360b,

title = "Multivariate receptor models and model uncertainty",

abstract = "Estimation of the number of major pollution sources, the source composition profiles, and the source contributions are the main interests in multivariate receptor modeling. Due to lack of identifiability of the receptor model, however, the estimation cannot be done without some additional assumptions. A common approach to this problem is to estimate the number of sources, q, at the first stage, and then estimate source profiles and contributions at the second stage, given additional constraints (identifiability conditions) to prevent source rotation/transformation and the assumption that the q-source model is correct. These assumptions on the parameters (the number of sources and identifiability conditions) are the main source of model uncertainty in multivariate receptor modeling. In this paper, we suggest a Bayesian approach to deal with model uncertainties in multivariate receptor models by using Markov chain Monte Carlo (MCMC) schemes. Specifically, we suggest a method which can simultaneously estimate parameters (compositions and contributions), parameter uncertainties, and model uncertainties (number of sources and identifiability conditions). Simulation results and an application to air pollution data are presented.",

keywords = "Factor analysis models, Latent variable models, Marginal likelihood, Model identifiability, Model uncertainty, Number of sources, Posterior model probability",

author = "Park, {Eun Sug} and Oh, {Man Suk} and Peter Guttorp",

note = "Funding Information: Although the research described in this article has been funded by the United States Environmental Protection Agency through agreement CR825173-01-0 to the University of Washington, it has not been subjected to the agency's required peer and policy review and therefore does not necessarily reflect the views of the Agency and no official endorsement should be inferred. ",

year = "2002",

month = jan,

day = "28",

doi = "10.1016/S0169-7439(01)00185-X",

language = "English",

volume = "60",

pages = "49--67",

journal = "Chemometrics and Intelligent Laboratory Systems",

issn = "0169-7439",

publisher = "Elsevier B.V.",

number = "1-2",

}

TY - JOUR

T1 - Multivariate receptor models and model uncertainty

AU - Park, Eun Sug

AU - Oh, Man Suk

AU - Guttorp, Peter

N1 - Funding Information: Although the research described in this article has been funded by the United States Environmental Protection Agency through agreement CR825173-01-0 to the University of Washington, it has not been subjected to the agency's required peer and policy review and therefore does not necessarily reflect the views of the Agency and no official endorsement should be inferred.

PY - 2002/1/28

Y1 - 2002/1/28

N2 - Estimation of the number of major pollution sources, the source composition profiles, and the source contributions are the main interests in multivariate receptor modeling. Due to lack of identifiability of the receptor model, however, the estimation cannot be done without some additional assumptions. A common approach to this problem is to estimate the number of sources, q, at the first stage, and then estimate source profiles and contributions at the second stage, given additional constraints (identifiability conditions) to prevent source rotation/transformation and the assumption that the q-source model is correct. These assumptions on the parameters (the number of sources and identifiability conditions) are the main source of model uncertainty in multivariate receptor modeling. In this paper, we suggest a Bayesian approach to deal with model uncertainties in multivariate receptor models by using Markov chain Monte Carlo (MCMC) schemes. Specifically, we suggest a method which can simultaneously estimate parameters (compositions and contributions), parameter uncertainties, and model uncertainties (number of sources and identifiability conditions). Simulation results and an application to air pollution data are presented.

AB - Estimation of the number of major pollution sources, the source composition profiles, and the source contributions are the main interests in multivariate receptor modeling. Due to lack of identifiability of the receptor model, however, the estimation cannot be done without some additional assumptions. A common approach to this problem is to estimate the number of sources, q, at the first stage, and then estimate source profiles and contributions at the second stage, given additional constraints (identifiability conditions) to prevent source rotation/transformation and the assumption that the q-source model is correct. These assumptions on the parameters (the number of sources and identifiability conditions) are the main source of model uncertainty in multivariate receptor modeling. In this paper, we suggest a Bayesian approach to deal with model uncertainties in multivariate receptor models by using Markov chain Monte Carlo (MCMC) schemes. Specifically, we suggest a method which can simultaneously estimate parameters (compositions and contributions), parameter uncertainties, and model uncertainties (number of sources and identifiability conditions). Simulation results and an application to air pollution data are presented.

KW - Factor analysis models

KW - Latent variable models

KW - Marginal likelihood

KW - Model identifiability

KW - Model uncertainty

KW - Number of sources

KW - Posterior model probability

UR - http://www.scopus.com/inward/record.url?scp=0037185404&partnerID=8YFLogxK

U2 - 10.1016/S0169-7439(01)00185-X

DO - 10.1016/S0169-7439(01)00185-X

M3 - Article

AN - SCOPUS:0037185404

SN - 0169-7439

VL - 60

SP - 49

EP - 67

JO - Chemometrics and Intelligent Laboratory Systems

JF - Chemometrics and Intelligent Laboratory Systems

IS - 1-2

ER -

Multivariate receptor models and model uncertainty

Abstract

Bibliographical note

Keywords

Access to Document

Fingerprint

Cite this