Classification of high-resolution NMR spectra based on complex wavelet domain feature selection and kernel-induced random forest

Guangzhe Fan, Zhou Wang, Seoung Bum Kim, Chivalai Temiyasathit

Research output: Chapter in Book/Report/Conference proceedingConference contribution

1 Citation (Scopus)

Abstract

High-resolution nuclear magnetic resonance (NMR) spectra contain important biomarkers that have potentials for early diagnosis of disease and subsequent monitoring of its progression. Traditional features extraction and analysis methods have been carried out in the original frequency spectrum domain. In this study, we conduct feature selection based on a complex wavelet transform by making use of its energy shift-insensitive property in a multi-resolution signal decomposition. A false discovery rate based multiple testing procedure is employed to identify important metabolite features. Furthermore, a novel kernel-induced random forest algorithm is used for the classification of NMR spectra based on the selected features. Our experiments with real NMR spectra showed that the proposed method leads to significant reduction in misclassification rate.

Original languageEnglish
Title of host publicationLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Pages593-600
Number of pages8
Volume6134 LNCS
DOIs
Publication statusPublished - 2010 Dec 1
Event4th International Conference on Image and Signal Processing, ICISP 2010 - Trois-Rivieres, QC, Canada
Duration: 2010 Jun 302010 Jul 2

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume6134 LNCS
ISSN (Print)03029743
ISSN (Electronic)16113349

Other

Other4th International Conference on Image and Signal Processing, ICISP 2010
CountryCanada
CityTrois-Rivieres, QC
Period10/6/3010/7/2

Fingerprint

Nuclear Magnetic Resonance
Random Forest
Feature Selection
Feature extraction
Wavelets
High Resolution
Nuclear magnetic resonance
kernel
Misclassification Rate
Multiple Testing
Frequency Spectrum
Biomarkers
Metabolites
Multiresolution
Progression
Wavelet transforms
Wavelet Transform
Feature Extraction
Monitoring
Decomposition

Keywords

  • Classification tree
  • Complex wavelet transforms
  • False discovery rate
  • High-resolution NMR spectrum
  • Kernel
  • Metabolomics
  • Random forest

ASJC Scopus subject areas

  • Computer Science(all)
  • Theoretical Computer Science

Cite this

Fan, G., Wang, Z., Kim, S. B., & Temiyasathit, C. (2010). Classification of high-resolution NMR spectra based on complex wavelet domain feature selection and kernel-induced random forest. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 6134 LNCS, pp. 593-600). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 6134 LNCS). https://doi.org/10.1007/978-3-642-13681-8_69

Classification of high-resolution NMR spectra based on complex wavelet domain feature selection and kernel-induced random forest. / Fan, Guangzhe; Wang, Zhou; Kim, Seoung Bum; Temiyasathit, Chivalai.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 6134 LNCS 2010. p. 593-600 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 6134 LNCS).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Fan, G, Wang, Z, Kim, SB & Temiyasathit, C 2010, Classification of high-resolution NMR spectra based on complex wavelet domain feature selection and kernel-induced random forest. in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). vol. 6134 LNCS, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 6134 LNCS, pp. 593-600, 4th International Conference on Image and Signal Processing, ICISP 2010, Trois-Rivieres, QC, Canada, 10/6/30. https://doi.org/10.1007/978-3-642-13681-8_69
Fan G, Wang Z, Kim SB, Temiyasathit C. Classification of high-resolution NMR spectra based on complex wavelet domain feature selection and kernel-induced random forest. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 6134 LNCS. 2010. p. 593-600. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-642-13681-8_69
Fan, Guangzhe ; Wang, Zhou ; Kim, Seoung Bum ; Temiyasathit, Chivalai. / Classification of high-resolution NMR spectra based on complex wavelet domain feature selection and kernel-induced random forest. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 6134 LNCS 2010. pp. 593-600 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{bd11fc871fb94915aa0c49734de4025c,
title = "Classification of high-resolution NMR spectra based on complex wavelet domain feature selection and kernel-induced random forest",
abstract = "High-resolution nuclear magnetic resonance (NMR) spectra contain important biomarkers that have potentials for early diagnosis of disease and subsequent monitoring of its progression. Traditional features extraction and analysis methods have been carried out in the original frequency spectrum domain. In this study, we conduct feature selection based on a complex wavelet transform by making use of its energy shift-insensitive property in a multi-resolution signal decomposition. A false discovery rate based multiple testing procedure is employed to identify important metabolite features. Furthermore, a novel kernel-induced random forest algorithm is used for the classification of NMR spectra based on the selected features. Our experiments with real NMR spectra showed that the proposed method leads to significant reduction in misclassification rate.",
keywords = "Classification tree, Complex wavelet transforms, False discovery rate, High-resolution NMR spectrum, Kernel, Metabolomics, Random forest",
author = "Guangzhe Fan and Zhou Wang and Kim, {Seoung Bum} and Chivalai Temiyasathit",
year = "2010",
month = "12",
day = "1",
doi = "10.1007/978-3-642-13681-8_69",
language = "English",
isbn = "364213680X",
volume = "6134 LNCS",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
pages = "593--600",
booktitle = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

}

TY - GEN

T1 - Classification of high-resolution NMR spectra based on complex wavelet domain feature selection and kernel-induced random forest

AU - Fan, Guangzhe

AU - Wang, Zhou

AU - Kim, Seoung Bum

AU - Temiyasathit, Chivalai

PY - 2010/12/1

Y1 - 2010/12/1

N2 - High-resolution nuclear magnetic resonance (NMR) spectra contain important biomarkers that have potentials for early diagnosis of disease and subsequent monitoring of its progression. Traditional features extraction and analysis methods have been carried out in the original frequency spectrum domain. In this study, we conduct feature selection based on a complex wavelet transform by making use of its energy shift-insensitive property in a multi-resolution signal decomposition. A false discovery rate based multiple testing procedure is employed to identify important metabolite features. Furthermore, a novel kernel-induced random forest algorithm is used for the classification of NMR spectra based on the selected features. Our experiments with real NMR spectra showed that the proposed method leads to significant reduction in misclassification rate.

AB - High-resolution nuclear magnetic resonance (NMR) spectra contain important biomarkers that have potentials for early diagnosis of disease and subsequent monitoring of its progression. Traditional features extraction and analysis methods have been carried out in the original frequency spectrum domain. In this study, we conduct feature selection based on a complex wavelet transform by making use of its energy shift-insensitive property in a multi-resolution signal decomposition. A false discovery rate based multiple testing procedure is employed to identify important metabolite features. Furthermore, a novel kernel-induced random forest algorithm is used for the classification of NMR spectra based on the selected features. Our experiments with real NMR spectra showed that the proposed method leads to significant reduction in misclassification rate.

KW - Classification tree

KW - Complex wavelet transforms

KW - False discovery rate

KW - High-resolution NMR spectrum

KW - Kernel

KW - Metabolomics

KW - Random forest

UR - http://www.scopus.com/inward/record.url?scp=79956288830&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=79956288830&partnerID=8YFLogxK

U2 - 10.1007/978-3-642-13681-8_69

DO - 10.1007/978-3-642-13681-8_69

M3 - Conference contribution

AN - SCOPUS:79956288830

SN - 364213680X

SN - 9783642136801

VL - 6134 LNCS

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 593

EP - 600

BT - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

ER -