Comparison of support vector machine classification to partial least squares dimension reduction with logistic discrimination of hyperspectral data

Machelle D. Wilson; Susan L. Ustin; David M Rocke

doi:10.1117/12.463169

Comparison of support vector machine classification to partial least squares dimension reduction with logistic discrimination of hyperspectral data

Machelle D. Wilson, Susan L. Ustin, David M Rocke

Public Health Sciences

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

1 Scopus citations

Abstract

The classification effectiveness of two relatively new techniques on data consisting of leaf-level reflectance from plants that have been exposed to varying levels of heavy metal toxicity was compared. The classification methods compared were support vector machine (SVM) classification of exposed and non-exposed plants based on the reflectance data, and partial least squares compression of the reflectance data followed by classification using logistics discrimination (PLS/LD). PLS/LD using binary predictor variables during compression had the lowest estimated prediction error for the data analyzed.

Original language	English (US)
Title of host publication	Proceedings of SPIE - The International Society for Optical Engineering
Editors	M. Ehlers
Pages	487-497
Number of pages	11
Volume	4886
DOIs	https://doi.org/10.1117/12.463169
State	Published - 2002
Event	Remote Sensing for Environmental Monitoring, GIS Applications, and Geology II - Agia Pelagia, Greece Duration: Sep 23 2002 → Sep 26 2002

Other

Other	Remote Sensing for Environmental Monitoring, GIS Applications, and Geology II
Country	Greece
City	Agia Pelagia
Period	9/23/02 → 9/26/02

Keywords

Heavy metals
Hyperspectral
Logistic discrimination
Partial least squares
Reflectance
Support vector machines

ASJC Scopus subject areas

Electrical and Electronic Engineering
Condensed Matter Physics

Access to Document

10.1117/12.463169

Fingerprint Dive into the research topics of 'Comparison of support vector machine classification to partial least squares dimension reduction with logistic discrimination of hyperspectral data'. Together they form a unique fingerprint.

Cite this

Wilson, M. D., Ustin, S. L., & Rocke, D. M. (2002). Comparison of support vector machine classification to partial least squares dimension reduction with logistic discrimination of hyperspectral data. In M. Ehlers (Ed.), Proceedings of SPIE - The International Society for Optical Engineering (Vol. 4886, pp. 487-497) https://doi.org/10.1117/12.463169

Comparison of support vector machine classification to partial least squares dimension reduction with logistic discrimination of hyperspectral data. / Wilson, Machelle D.; Ustin, Susan L.; Rocke, David M.

Proceedings of SPIE - The International Society for Optical Engineering. ed. / M. Ehlers. Vol. 4886 2002. p. 487-497.

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Wilson, MD, Ustin, SL & Rocke, DM 2002, Comparison of support vector machine classification to partial least squares dimension reduction with logistic discrimination of hyperspectral data. in M Ehlers (ed.), Proceedings of SPIE - The International Society for Optical Engineering. vol. 4886, pp. 487-497, Remote Sensing for Environmental Monitoring, GIS Applications, and Geology II, Agia Pelagia, Greece, 9/23/02. https://doi.org/10.1117/12.463169

@inproceedings{1b238575486345f6960d168b8feb539d,

title = "Comparison of support vector machine classification to partial least squares dimension reduction with logistic discrimination of hyperspectral data",

abstract = "The classification effectiveness of two relatively new techniques on data consisting of leaf-level reflectance from plants that have been exposed to varying levels of heavy metal toxicity was compared. The classification methods compared were support vector machine (SVM) classification of exposed and non-exposed plants based on the reflectance data, and partial least squares compression of the reflectance data followed by classification using logistics discrimination (PLS/LD). PLS/LD using binary predictor variables during compression had the lowest estimated prediction error for the data analyzed.",

keywords = "Heavy metals, Hyperspectral, Logistic discrimination, Partial least squares, Reflectance, Support vector machines",

author = "Wilson, {Machelle D.} and Ustin, {Susan L.} and Rocke, {David M}",

year = "2002",

doi = "10.1117/12.463169",

language = "English (US)",

volume = "4886",

pages = "487--497",

editor = "M. Ehlers",

booktitle = "Proceedings of SPIE - The International Society for Optical Engineering",

note = "Remote Sensing for Environmental Monitoring, GIS Applications, and Geology II ; Conference date: 23-09-2002 Through 26-09-2002",

}

TY - GEN

T1 - Comparison of support vector machine classification to partial least squares dimension reduction with logistic discrimination of hyperspectral data

AU - Wilson, Machelle D.

AU - Ustin, Susan L.

AU - Rocke, David M

PY - 2002

Y1 - 2002

N2 - The classification effectiveness of two relatively new techniques on data consisting of leaf-level reflectance from plants that have been exposed to varying levels of heavy metal toxicity was compared. The classification methods compared were support vector machine (SVM) classification of exposed and non-exposed plants based on the reflectance data, and partial least squares compression of the reflectance data followed by classification using logistics discrimination (PLS/LD). PLS/LD using binary predictor variables during compression had the lowest estimated prediction error for the data analyzed.

AB - The classification effectiveness of two relatively new techniques on data consisting of leaf-level reflectance from plants that have been exposed to varying levels of heavy metal toxicity was compared. The classification methods compared were support vector machine (SVM) classification of exposed and non-exposed plants based on the reflectance data, and partial least squares compression of the reflectance data followed by classification using logistics discrimination (PLS/LD). PLS/LD using binary predictor variables during compression had the lowest estimated prediction error for the data analyzed.

KW - Heavy metals

KW - Hyperspectral

KW - Logistic discrimination

KW - Partial least squares

KW - Reflectance

KW - Support vector machines

UR - http://www.scopus.com/inward/record.url?scp=0037951522&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0037951522&partnerID=8YFLogxK

U2 - 10.1117/12.463169

DO - 10.1117/12.463169

M3 - Conference contribution

AN - SCOPUS:0037951522

VL - 4886

SP - 487

EP - 497

BT - Proceedings of SPIE - The International Society for Optical Engineering

A2 - Ehlers, M.

T2 - Remote Sensing for Environmental Monitoring, GIS Applications, and Geology II

Y2 - 23 September 2002 through 26 September 2002

ER -