Data reduction using a discrete wavelet transform in discriminant analysis of very high dimensionality data

Yinsheng Qu; Bao Ling Adam; Mark Thornquist; John D. Potter; Mary Lou Thompson; Yutaka Yasui; John Davis; Paul F. Schellhammer; Lisa Cazares; Mary Ann Clements; George L. Wright; Ziding Feng

doi:10.1111/1541-0420.00017

Data reduction using a discrete wavelet transform in discriminant analysis of very high dimensionality data

Yinsheng Qu, Bao Ling Adam, Mark Thornquist, John D. Potter, Mary Lou Thompson, Yutaka Yasui, John Davis, Paul F. Schellhammer, Lisa Cazares, Mary Ann Clements, George L. Wright, Ziding Feng

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

63 Scopus citations

Abstract

We present a method of data reduction using a wavelet transform in discriminant analysis when the number of variables is much greater than the number of observations. The method is illustrated with a prostate cancer study, where the sample size is 248, and the number of variables is 48,538 (generated using the ProteinChip technology). Using a discrete wavelet transform, the 48,538 data points are represented by 1271 wavelet coefficients. Information criteria identified 11 of the 1271 wavelet coefficients with the highest discriminatory power. The linear classifier with the 11 wavelet coefficients detected prostate cancer in a separate test set with a sensitivity of 97% and specificity of 100%.

Original language	English (US)
Pages (from-to)	143-151
Number of pages	9
Journal	Biometrics
Volume	59
Issue number	1
DOIs	https://doi.org/10.1111/1541-0420.00017
State	Published - Mar 2003
Externally published	Yes

Keywords

Area under the ROC curve
Divergence
Fisher discriminant analysis
Kullback-Leibler information
Mahalanobis distance
Principal components analysis

ASJC Scopus subject areas

Statistics and Probability
General Biochemistry, Genetics and Molecular Biology
General Immunology and Microbiology
General Agricultural and Biological Sciences
Applied Mathematics

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

Access to Document

10.1111/1541-0420.00017

Cite this

@article{2e1c7f30809d49df9b94a81747ab4262,

title = "Data reduction using a discrete wavelet transform in discriminant analysis of very high dimensionality data",

abstract = "We present a method of data reduction using a wavelet transform in discriminant analysis when the number of variables is much greater than the number of observations. The method is illustrated with a prostate cancer study, where the sample size is 248, and the number of variables is 48,538 (generated using the ProteinChip technology). Using a discrete wavelet transform, the 48,538 data points are represented by 1271 wavelet coefficients. Information criteria identified 11 of the 1271 wavelet coefficients with the highest discriminatory power. The linear classifier with the 11 wavelet coefficients detected prostate cancer in a separate test set with a sensitivity of 97% and specificity of 100%.",

keywords = "Area under the ROC curve, Divergence, Fisher discriminant analysis, Kullback-Leibler information, Mahalanobis distance, Principal components analysis",

author = "Yinsheng Qu and Adam, {Bao Ling} and Mark Thornquist and Potter, {John D.} and Thompson, {Mary Lou} and Yutaka Yasui and John Davis and Schellhammer, {Paul F.} and Lisa Cazares and Clements, {Mary Ann} and Wright, {George L.} and Ziding Feng",

year = "2003",

month = mar,

doi = "10.1111/1541-0420.00017",

language = "English (US)",

volume = "59",

pages = "143--151",

journal = "Biometrics",

issn = "0006-341X",

publisher = "Wiley-Blackwell",

number = "1",

}

TY - JOUR

T1 - Data reduction using a discrete wavelet transform in discriminant analysis of very high dimensionality data

AU - Qu, Yinsheng

AU - Adam, Bao Ling

AU - Thornquist, Mark

AU - Potter, John D.

AU - Thompson, Mary Lou

AU - Yasui, Yutaka

AU - Davis, John

AU - Schellhammer, Paul F.

AU - Cazares, Lisa

AU - Clements, Mary Ann

AU - Wright, George L.

AU - Feng, Ziding

PY - 2003/3

Y1 - 2003/3

N2 - We present a method of data reduction using a wavelet transform in discriminant analysis when the number of variables is much greater than the number of observations. The method is illustrated with a prostate cancer study, where the sample size is 248, and the number of variables is 48,538 (generated using the ProteinChip technology). Using a discrete wavelet transform, the 48,538 data points are represented by 1271 wavelet coefficients. Information criteria identified 11 of the 1271 wavelet coefficients with the highest discriminatory power. The linear classifier with the 11 wavelet coefficients detected prostate cancer in a separate test set with a sensitivity of 97% and specificity of 100%.

AB - We present a method of data reduction using a wavelet transform in discriminant analysis when the number of variables is much greater than the number of observations. The method is illustrated with a prostate cancer study, where the sample size is 248, and the number of variables is 48,538 (generated using the ProteinChip technology). Using a discrete wavelet transform, the 48,538 data points are represented by 1271 wavelet coefficients. Information criteria identified 11 of the 1271 wavelet coefficients with the highest discriminatory power. The linear classifier with the 11 wavelet coefficients detected prostate cancer in a separate test set with a sensitivity of 97% and specificity of 100%.

KW - Area under the ROC curve

KW - Divergence

KW - Fisher discriminant analysis

KW - Kullback-Leibler information

KW - Mahalanobis distance

KW - Principal components analysis

UR - http://www.scopus.com/inward/record.url?scp=0242499870&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0242499870&partnerID=8YFLogxK

U2 - 10.1111/1541-0420.00017

DO - 10.1111/1541-0420.00017

M3 - Article

C2 - 12762451

AN - SCOPUS:0242499870

SN - 0006-341X

VL - 59

SP - 143

EP - 151

JO - Biometrics

JF - Biometrics

IS - 1

ER -

Data reduction using a discrete wavelet transform in discriminant analysis of very high dimensionality data

Abstract

Keywords

ASJC Scopus subject areas

UN SDGs

Access to Document

Other files and links

Fingerprint

Cite this