A monotonic measure for optimal feature selection

Huan Liu, Hiroshi Motoda, Manoranjan Dash

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

38 Citations (Scopus)

Abstract

Feature selection is a problem of choosing a subset of relevant features. In general,only exhaustive search can bring about the optimal subset. With a monotonic measure, exhaustive search can be avoided without sacrificing optimality. Unfortunately, most error- or distance-based measures are not monotonic. A new measure is employed in this work that is monotonic and fast to compute. The search for relevant features according to this measure is guaranteed tobe complete but not exhaustive. Experiments are conducted for verification.

Original languageEnglish (US)
Title of host publicationLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
PublisherSpringer Verlag
Pages101-106
Number of pages6
Volume1398
ISBN (Print)3540644172, 9783540644170
StatePublished - 1998
Externally publishedYes
Event10th European Conference on Machine Learning, ECML 1998 - Chemnitz, Germany
Duration: Apr 21 1998Apr 23 1998

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume1398
ISSN (Print)03029743
ISSN (Electronic)16113349

Other

Other10th European Conference on Machine Learning, ECML 1998
CountryGermany
CityChemnitz
Period4/21/984/23/98

Fingerprint

Monotonic
Feature Selection
Feature extraction
Exhaustive Search
Experiments
Subset
Optimality
Experiment

ASJC Scopus subject areas

  • Computer Science(all)
  • Theoretical Computer Science

Cite this

Liu, H., Motoda, H., & Dash, M. (1998). A monotonic measure for optimal feature selection. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 1398, pp. 101-106). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 1398). Springer Verlag.

A monotonic measure for optimal feature selection. / Liu, Huan; Motoda, Hiroshi; Dash, Manoranjan.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 1398 Springer Verlag, 1998. p. 101-106 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 1398).

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

Liu, H, Motoda, H & Dash, M 1998, A monotonic measure for optimal feature selection. in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). vol. 1398, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 1398, Springer Verlag, pp. 101-106, 10th European Conference on Machine Learning, ECML 1998, Chemnitz, Germany, 4/21/98.
Liu H, Motoda H, Dash M. A monotonic measure for optimal feature selection. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 1398. Springer Verlag. 1998. p. 101-106. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
Liu, Huan ; Motoda, Hiroshi ; Dash, Manoranjan. / A monotonic measure for optimal feature selection. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 1398 Springer Verlag, 1998. pp. 101-106 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{00730a304d864d12913a83cb88122064,
title = "A monotonic measure for optimal feature selection",
abstract = "Feature selection is a problem of choosing a subset of relevant features. In general,only exhaustive search can bring about the optimal subset. With a monotonic measure, exhaustive search can be avoided without sacrificing optimality. Unfortunately, most error- or distance-based measures are not monotonic. A new measure is employed in this work that is monotonic and fast to compute. The search for relevant features according to this measure is guaranteed tobe complete but not exhaustive. Experiments are conducted for verification.",
author = "Huan Liu and Hiroshi Motoda and Manoranjan Dash",
year = "1998",
language = "English (US)",
isbn = "3540644172",
volume = "1398",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "101--106",
booktitle = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

}

TY - GEN

T1 - A monotonic measure for optimal feature selection

AU - Liu, Huan

AU - Motoda, Hiroshi

AU - Dash, Manoranjan

PY - 1998

Y1 - 1998

N2 - Feature selection is a problem of choosing a subset of relevant features. In general,only exhaustive search can bring about the optimal subset. With a monotonic measure, exhaustive search can be avoided without sacrificing optimality. Unfortunately, most error- or distance-based measures are not monotonic. A new measure is employed in this work that is monotonic and fast to compute. The search for relevant features according to this measure is guaranteed tobe complete but not exhaustive. Experiments are conducted for verification.

AB - Feature selection is a problem of choosing a subset of relevant features. In general,only exhaustive search can bring about the optimal subset. With a monotonic measure, exhaustive search can be avoided without sacrificing optimality. Unfortunately, most error- or distance-based measures are not monotonic. A new measure is employed in this work that is monotonic and fast to compute. The search for relevant features according to this measure is guaranteed tobe complete but not exhaustive. Experiments are conducted for verification.

UR - http://www.scopus.com/inward/record.url?scp=79954993774&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=79954993774&partnerID=8YFLogxK

M3 - Conference contribution

SN - 3540644172

SN - 9783540644170

VL - 1398

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

SP - 101

EP - 106

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

PB - Springer Verlag

ER -