Hierarchical multi-label classification with SVMs: A case study in gene function prediction

Peerapon Vateekul, Miroslav Kubat, Kanoksri Sarinnapakorn

Research output: Contribution to journalArticle

12 Citations (Scopus)

Abstract

Hierarchical multi-label classification is a relatively new research topic in the field of classifier induction. What distinguishes it from earlier tasks is that it allows each example to belong to two or more classes at the same time, and by assuming that the classes are mutually related by generalization/specialization operators. The paper first investigates the problem of performance evaluation in these domains. After this, it proposes a new induction system, HR-SVM, built around support vector machines. In our experiments, we demonstrate that this system's performance compares favorably with that earlier attempts, and then we proceed to an investigation of how HR-SVM's individual modules contribute to the overall system's behavior. As a testbed, we use a set of benchmark domains from the field of gene-function prediction.

Original languageEnglish (US)
Pages (from-to)717-738
Number of pages22
JournalIntelligent Data Analysis
Volume18
Issue number4
DOIs
StatePublished - 2014

Fingerprint

Testbeds
Support vector machines
Labels
Proof by induction
Classifiers
Genes
Gene
Prediction
Specialization
Testbed
Performance Evaluation
System Performance
Support Vector Machine
Experiments
Classifier
Benchmark
Module
Operator
Demonstrate
Experiment

Keywords

  • gene-function prediction
  • Hierarchical multi-label classification
  • support vector machines

ASJC Scopus subject areas

  • Artificial Intelligence
  • Theoretical Computer Science
  • Computer Vision and Pattern Recognition

Cite this

Hierarchical multi-label classification with SVMs : A case study in gene function prediction. / Vateekul, Peerapon; Kubat, Miroslav; Sarinnapakorn, Kanoksri.

In: Intelligent Data Analysis, Vol. 18, No. 4, 2014, p. 717-738.

Research output: Contribution to journalArticle

Vateekul, Peerapon ; Kubat, Miroslav ; Sarinnapakorn, Kanoksri. / Hierarchical multi-label classification with SVMs : A case study in gene function prediction. In: Intelligent Data Analysis. 2014 ; Vol. 18, No. 4. pp. 717-738.
@article{cbd9ebf5419e4e1486e1231bcb04fc9e,
title = "Hierarchical multi-label classification with SVMs: A case study in gene function prediction",
abstract = "Hierarchical multi-label classification is a relatively new research topic in the field of classifier induction. What distinguishes it from earlier tasks is that it allows each example to belong to two or more classes at the same time, and by assuming that the classes are mutually related by generalization/specialization operators. The paper first investigates the problem of performance evaluation in these domains. After this, it proposes a new induction system, HR-SVM, built around support vector machines. In our experiments, we demonstrate that this system's performance compares favorably with that earlier attempts, and then we proceed to an investigation of how HR-SVM's individual modules contribute to the overall system's behavior. As a testbed, we use a set of benchmark domains from the field of gene-function prediction.",
keywords = "gene-function prediction, Hierarchical multi-label classification, support vector machines",
author = "Peerapon Vateekul and Miroslav Kubat and Kanoksri Sarinnapakorn",
year = "2014",
doi = "10.3233/IDA-140665",
language = "English (US)",
volume = "18",
pages = "717--738",
journal = "Intelligent Data Analysis",
issn = "1088-467X",
publisher = "IOS Press",
number = "4",

}

TY - JOUR

T1 - Hierarchical multi-label classification with SVMs

T2 - A case study in gene function prediction

AU - Vateekul, Peerapon

AU - Kubat, Miroslav

AU - Sarinnapakorn, Kanoksri

PY - 2014

Y1 - 2014

N2 - Hierarchical multi-label classification is a relatively new research topic in the field of classifier induction. What distinguishes it from earlier tasks is that it allows each example to belong to two or more classes at the same time, and by assuming that the classes are mutually related by generalization/specialization operators. The paper first investigates the problem of performance evaluation in these domains. After this, it proposes a new induction system, HR-SVM, built around support vector machines. In our experiments, we demonstrate that this system's performance compares favorably with that earlier attempts, and then we proceed to an investigation of how HR-SVM's individual modules contribute to the overall system's behavior. As a testbed, we use a set of benchmark domains from the field of gene-function prediction.

AB - Hierarchical multi-label classification is a relatively new research topic in the field of classifier induction. What distinguishes it from earlier tasks is that it allows each example to belong to two or more classes at the same time, and by assuming that the classes are mutually related by generalization/specialization operators. The paper first investigates the problem of performance evaluation in these domains. After this, it proposes a new induction system, HR-SVM, built around support vector machines. In our experiments, we demonstrate that this system's performance compares favorably with that earlier attempts, and then we proceed to an investigation of how HR-SVM's individual modules contribute to the overall system's behavior. As a testbed, we use a set of benchmark domains from the field of gene-function prediction.

KW - gene-function prediction

KW - Hierarchical multi-label classification

KW - support vector machines

UR - http://www.scopus.com/inward/record.url?scp=84948128179&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84948128179&partnerID=8YFLogxK

U2 - 10.3233/IDA-140665

DO - 10.3233/IDA-140665

M3 - Article

AN - SCOPUS:84948128179

VL - 18

SP - 717

EP - 738

JO - Intelligent Data Analysis

JF - Intelligent Data Analysis

SN - 1088-467X

IS - 4

ER -