A comparative study of feature selection and multiclass classfication methods for tissue classification based on gene expression

Tao Li, Chengliang Zhang, Mitsunori Ogihara

Research output: Contribution to journalArticle

450 Citations (Scopus)

Abstract

Summary: This paper studies the problem of building multiclass classifiers for tissue classification based on gene expression. The recent development of microarray technologies has enabled biologists to quantify gene expression of tens of thousands of genes in a single experiment. Biologists have begun collecting gene expression for a large number of samples. One of the urgent issues in the use of microarray data is to develop methods for characterizing samples based on their gene expression. The most basic step in the research direction is binary sample classification, which has been studied extensively over the past few years. This paper investigates the next step - multiclass classification of samples based on gene expression. The characteristics of expression data (e.g. large number of genes with small sample size) makes the classification problem more challenging. The process of building multiclass classifiers is divided into two components: (i) selection of the features (i.e. genes) to be used for training and testing and (ii) selection of the classification method. This paper compares various feature selection methods as well as various state-of-the-art classification methods on various multiclass gene expression datasets. Our study indicates that multiclass classification problem is much more difficult than the binary one for the gene expression datasets. The difficulty lies in the fact that the data are of high dimensionality and that the sample size is small. The classification accuracy appears to degrade very rapidly as the number of classes increases. In particular, the accuracy was very low regardless of the choices of the methods for large-class datasets (e.g. NCI60 and GCM). While increasing the number of samples is a plausible solution to the problem of accuracy degradation, it is important to develop algorithms that are able to analyze effectively multiple-class expression data for these special datasets.

Original languageEnglish (US)
Pages (from-to)2429-2437
Number of pages9
JournalBioinformatics
Volume20
Issue number15
DOIs
StatePublished - Oct 12 2004
Externally publishedYes

Fingerprint

Multi-class
Gene expression
Feature Selection
Gene Expression
Comparative Study
Feature extraction
Tissue
Multi-class Classification
Gene
Genes
Classification Problems
Microarrays
Sample Size
Classifier
Classifiers
Binary
Large Data
Small Sample Size
Microarray Data
Microarray

ASJC Scopus subject areas

  • Clinical Biochemistry
  • Computer Science Applications
  • Computational Theory and Mathematics

Cite this

A comparative study of feature selection and multiclass classfication methods for tissue classification based on gene expression. / Li, Tao; Zhang, Chengliang; Ogihara, Mitsunori.

In: Bioinformatics, Vol. 20, No. 15, 12.10.2004, p. 2429-2437.

Research output: Contribution to journalArticle

@article{de86ebcb010e473ab74c09c0cb13ed24,
title = "A comparative study of feature selection and multiclass classfication methods for tissue classification based on gene expression",
abstract = "Summary: This paper studies the problem of building multiclass classifiers for tissue classification based on gene expression. The recent development of microarray technologies has enabled biologists to quantify gene expression of tens of thousands of genes in a single experiment. Biologists have begun collecting gene expression for a large number of samples. One of the urgent issues in the use of microarray data is to develop methods for characterizing samples based on their gene expression. The most basic step in the research direction is binary sample classification, which has been studied extensively over the past few years. This paper investigates the next step - multiclass classification of samples based on gene expression. The characteristics of expression data (e.g. large number of genes with small sample size) makes the classification problem more challenging. The process of building multiclass classifiers is divided into two components: (i) selection of the features (i.e. genes) to be used for training and testing and (ii) selection of the classification method. This paper compares various feature selection methods as well as various state-of-the-art classification methods on various multiclass gene expression datasets. Our study indicates that multiclass classification problem is much more difficult than the binary one for the gene expression datasets. The difficulty lies in the fact that the data are of high dimensionality and that the sample size is small. The classification accuracy appears to degrade very rapidly as the number of classes increases. In particular, the accuracy was very low regardless of the choices of the methods for large-class datasets (e.g. NCI60 and GCM). While increasing the number of samples is a plausible solution to the problem of accuracy degradation, it is important to develop algorithms that are able to analyze effectively multiple-class expression data for these special datasets.",
author = "Tao Li and Chengliang Zhang and Mitsunori Ogihara",
year = "2004",
month = "10",
day = "12",
doi = "10.1093/bioinformatics/bth267",
language = "English (US)",
volume = "20",
pages = "2429--2437",
journal = "Bioinformatics (Oxford, England)",
issn = "1367-4803",
publisher = "Oxford University Press",
number = "15",

}

TY - JOUR

T1 - A comparative study of feature selection and multiclass classfication methods for tissue classification based on gene expression

AU - Li, Tao

AU - Zhang, Chengliang

AU - Ogihara, Mitsunori

PY - 2004/10/12

Y1 - 2004/10/12

N2 - Summary: This paper studies the problem of building multiclass classifiers for tissue classification based on gene expression. The recent development of microarray technologies has enabled biologists to quantify gene expression of tens of thousands of genes in a single experiment. Biologists have begun collecting gene expression for a large number of samples. One of the urgent issues in the use of microarray data is to develop methods for characterizing samples based on their gene expression. The most basic step in the research direction is binary sample classification, which has been studied extensively over the past few years. This paper investigates the next step - multiclass classification of samples based on gene expression. The characteristics of expression data (e.g. large number of genes with small sample size) makes the classification problem more challenging. The process of building multiclass classifiers is divided into two components: (i) selection of the features (i.e. genes) to be used for training and testing and (ii) selection of the classification method. This paper compares various feature selection methods as well as various state-of-the-art classification methods on various multiclass gene expression datasets. Our study indicates that multiclass classification problem is much more difficult than the binary one for the gene expression datasets. The difficulty lies in the fact that the data are of high dimensionality and that the sample size is small. The classification accuracy appears to degrade very rapidly as the number of classes increases. In particular, the accuracy was very low regardless of the choices of the methods for large-class datasets (e.g. NCI60 and GCM). While increasing the number of samples is a plausible solution to the problem of accuracy degradation, it is important to develop algorithms that are able to analyze effectively multiple-class expression data for these special datasets.

AB - Summary: This paper studies the problem of building multiclass classifiers for tissue classification based on gene expression. The recent development of microarray technologies has enabled biologists to quantify gene expression of tens of thousands of genes in a single experiment. Biologists have begun collecting gene expression for a large number of samples. One of the urgent issues in the use of microarray data is to develop methods for characterizing samples based on their gene expression. The most basic step in the research direction is binary sample classification, which has been studied extensively over the past few years. This paper investigates the next step - multiclass classification of samples based on gene expression. The characteristics of expression data (e.g. large number of genes with small sample size) makes the classification problem more challenging. The process of building multiclass classifiers is divided into two components: (i) selection of the features (i.e. genes) to be used for training and testing and (ii) selection of the classification method. This paper compares various feature selection methods as well as various state-of-the-art classification methods on various multiclass gene expression datasets. Our study indicates that multiclass classification problem is much more difficult than the binary one for the gene expression datasets. The difficulty lies in the fact that the data are of high dimensionality and that the sample size is small. The classification accuracy appears to degrade very rapidly as the number of classes increases. In particular, the accuracy was very low regardless of the choices of the methods for large-class datasets (e.g. NCI60 and GCM). While increasing the number of samples is a plausible solution to the problem of accuracy degradation, it is important to develop algorithms that are able to analyze effectively multiple-class expression data for these special datasets.

UR - http://www.scopus.com/inward/record.url?scp=7244248755&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=7244248755&partnerID=8YFLogxK

U2 - 10.1093/bioinformatics/bth267

DO - 10.1093/bioinformatics/bth267

M3 - Article

C2 - 15087314

AN - SCOPUS:7244248755

VL - 20

SP - 2429

EP - 2437

JO - Bioinformatics (Oxford, England)

JF - Bioinformatics (Oxford, England)

SN - 1367-4803

IS - 15

ER -