A novel peak-hopping stepwise feature selection method with application to Raman spectroscopy

Michael J. McShane, Brent D. Cameron, Gerard L. Coté, Massoud Motamedi, Clifford H. Spiegelman

Research output: Contribution to journalArticle

26 Citations (Scopus)

Abstract

A new stepwise approach to variable selection for spectroscopy that includes chemical information and attempts to test several spectral regions producing high ranking coefficients has been developed to improve on currently available methods. Existing selection techniques can, in general, be placed into two groups: the first, time-consuming optimization approaches that ignore available information about sample chemistry and require considerable expertise to arrive at appropriate solutions (e.g. genetic algorithms), and the second, stepwise procedures that tend to select many variables in the same area containing redundant information. The algorithm described here is a fast stepwise procedure that uses multiple ranking chains to identify several spectral regions correlated with known sample properties. The multiple-chain approach allows the generation of a final ranking vector that moves quickly away from the initial selection point, testing several areas exhibiting correlation between spectra and composition early in the stepping procedure. Quantitative evidence of the success of this approach as applied to Raman spectroscopy is given in terms of processing speed, number of selected variables, and prediction error in comparison with other selection methods. In this respect, the procedure described here may be considered as a significant evolutionary step in variable selection algorithms. Copyright (C) 1999 Elsevier Science B.V.

Original languageEnglish (US)
Pages (from-to)251-264
Number of pages14
JournalAnalytica Chimica Acta
Volume388
Issue number3
DOIs
StatePublished - May 7 1999
Externally publishedYes

Fingerprint

Raman Spectrum Analysis
Raman spectroscopy
Feature extraction
ranking
Genetic algorithms
Spectroscopy
Testing
Processing
Chemical analysis
multiple use
genetic algorithm
spectroscopy
method
Spectrum Analysis
prediction

Keywords

  • Analytical chemistry
  • Chemometrics
  • Multivariate calibration
  • Optimization
  • Partial least-squares
  • Wavelength selection

ASJC Scopus subject areas

  • Biochemistry
  • Analytical Chemistry
  • Spectroscopy
  • Environmental Chemistry

Cite this

A novel peak-hopping stepwise feature selection method with application to Raman spectroscopy. / McShane, Michael J.; Cameron, Brent D.; Coté, Gerard L.; Motamedi, Massoud; Spiegelman, Clifford H.

In: Analytica Chimica Acta, Vol. 388, No. 3, 07.05.1999, p. 251-264.

Research output: Contribution to journalArticle

McShane, Michael J. ; Cameron, Brent D. ; Coté, Gerard L. ; Motamedi, Massoud ; Spiegelman, Clifford H. / A novel peak-hopping stepwise feature selection method with application to Raman spectroscopy. In: Analytica Chimica Acta. 1999 ; Vol. 388, No. 3. pp. 251-264.
@article{021753ce3f834976b2eaab38620fde7e,
title = "A novel peak-hopping stepwise feature selection method with application to Raman spectroscopy",
abstract = "A new stepwise approach to variable selection for spectroscopy that includes chemical information and attempts to test several spectral regions producing high ranking coefficients has been developed to improve on currently available methods. Existing selection techniques can, in general, be placed into two groups: the first, time-consuming optimization approaches that ignore available information about sample chemistry and require considerable expertise to arrive at appropriate solutions (e.g. genetic algorithms), and the second, stepwise procedures that tend to select many variables in the same area containing redundant information. The algorithm described here is a fast stepwise procedure that uses multiple ranking chains to identify several spectral regions correlated with known sample properties. The multiple-chain approach allows the generation of a final ranking vector that moves quickly away from the initial selection point, testing several areas exhibiting correlation between spectra and composition early in the stepping procedure. Quantitative evidence of the success of this approach as applied to Raman spectroscopy is given in terms of processing speed, number of selected variables, and prediction error in comparison with other selection methods. In this respect, the procedure described here may be considered as a significant evolutionary step in variable selection algorithms. Copyright (C) 1999 Elsevier Science B.V.",
keywords = "Analytical chemistry, Chemometrics, Multivariate calibration, Optimization, Partial least-squares, Wavelength selection",
author = "McShane, {Michael J.} and Cameron, {Brent D.} and Cot{\'e}, {Gerard L.} and Massoud Motamedi and Spiegelman, {Clifford H.}",
year = "1999",
month = "5",
day = "7",
doi = "10.1016/S0003-2670(99)00080-X",
language = "English (US)",
volume = "388",
pages = "251--264",
journal = "Analytica Chimica Acta",
issn = "0003-2670",
publisher = "Elsevier",
number = "3",

}

TY - JOUR

T1 - A novel peak-hopping stepwise feature selection method with application to Raman spectroscopy

AU - McShane, Michael J.

AU - Cameron, Brent D.

AU - Coté, Gerard L.

AU - Motamedi, Massoud

AU - Spiegelman, Clifford H.

PY - 1999/5/7

Y1 - 1999/5/7

N2 - A new stepwise approach to variable selection for spectroscopy that includes chemical information and attempts to test several spectral regions producing high ranking coefficients has been developed to improve on currently available methods. Existing selection techniques can, in general, be placed into two groups: the first, time-consuming optimization approaches that ignore available information about sample chemistry and require considerable expertise to arrive at appropriate solutions (e.g. genetic algorithms), and the second, stepwise procedures that tend to select many variables in the same area containing redundant information. The algorithm described here is a fast stepwise procedure that uses multiple ranking chains to identify several spectral regions correlated with known sample properties. The multiple-chain approach allows the generation of a final ranking vector that moves quickly away from the initial selection point, testing several areas exhibiting correlation between spectra and composition early in the stepping procedure. Quantitative evidence of the success of this approach as applied to Raman spectroscopy is given in terms of processing speed, number of selected variables, and prediction error in comparison with other selection methods. In this respect, the procedure described here may be considered as a significant evolutionary step in variable selection algorithms. Copyright (C) 1999 Elsevier Science B.V.

AB - A new stepwise approach to variable selection for spectroscopy that includes chemical information and attempts to test several spectral regions producing high ranking coefficients has been developed to improve on currently available methods. Existing selection techniques can, in general, be placed into two groups: the first, time-consuming optimization approaches that ignore available information about sample chemistry and require considerable expertise to arrive at appropriate solutions (e.g. genetic algorithms), and the second, stepwise procedures that tend to select many variables in the same area containing redundant information. The algorithm described here is a fast stepwise procedure that uses multiple ranking chains to identify several spectral regions correlated with known sample properties. The multiple-chain approach allows the generation of a final ranking vector that moves quickly away from the initial selection point, testing several areas exhibiting correlation between spectra and composition early in the stepping procedure. Quantitative evidence of the success of this approach as applied to Raman spectroscopy is given in terms of processing speed, number of selected variables, and prediction error in comparison with other selection methods. In this respect, the procedure described here may be considered as a significant evolutionary step in variable selection algorithms. Copyright (C) 1999 Elsevier Science B.V.

KW - Analytical chemistry

KW - Chemometrics

KW - Multivariate calibration

KW - Optimization

KW - Partial least-squares

KW - Wavelength selection

UR - http://www.scopus.com/inward/record.url?scp=0032965618&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0032965618&partnerID=8YFLogxK

U2 - 10.1016/S0003-2670(99)00080-X

DO - 10.1016/S0003-2670(99)00080-X

M3 - Article

AN - SCOPUS:0032965618

VL - 388

SP - 251

EP - 264

JO - Analytica Chimica Acta

JF - Analytica Chimica Acta

SN - 0003-2670

IS - 3

ER -