An Exponential Monte-Carlo algorithm for feature selection problems

Research output: Contribution to journalArticle

11 Citations (Scopus)

Abstract

Feature selection problems (FS) can be defined as the process of eliminating redundant features while avoiding information loss. Due to that fact that FS is an NP-hard problem, heuristic and meta-heuristic approaches have been widely used by researchers. In this work, we proposed an Exponential Monte-Carlo algorithm (EMC-FS) for the feature selection problem. EMC-FS is a meta-heuristic approach which is quite similar to a simulated annealing algorithm. The difference is that no cooling schedule is required. Improved solutions are accepted and worse solutions are adaptively accepted based on the quality of the trial solution, the search time and the number of consecutive non-improving iterations. We have evaluated our approach against the latest methodologies in the literature on standard benchmark problems. The quality of the obtained subset of features has also been evaluated in terms of the number of generated rules (descriptive patterns) and classification accuracy. Our research demonstrates that our approach produces some of the best known results.

Original languageEnglish
Pages (from-to)160-167
Number of pages8
JournalComputers and Industrial Engineering
Volume67
Issue number1
DOIs
Publication statusPublished - Jan 2014

Fingerprint

Feature extraction
Electromagnetic compatibility
Simulated annealing
Computational complexity
Cooling

Keywords

  • Exponential Monte-Carlo
  • Feature selection
  • Local search

ASJC Scopus subject areas

  • Computer Science(all)
  • Engineering(all)

Cite this

An Exponential Monte-Carlo algorithm for feature selection problems. / Abdullah, Salwani; Sabar, Nasser R.; Ahmad Nazri, Mohd Zakree; Ayob, Masri.

In: Computers and Industrial Engineering, Vol. 67, No. 1, 01.2014, p. 160-167.

Research output: Contribution to journalArticle

@article{233944b6e2d6424091fb958ca0681b35,
title = "An Exponential Monte-Carlo algorithm for feature selection problems",
abstract = "Feature selection problems (FS) can be defined as the process of eliminating redundant features while avoiding information loss. Due to that fact that FS is an NP-hard problem, heuristic and meta-heuristic approaches have been widely used by researchers. In this work, we proposed an Exponential Monte-Carlo algorithm (EMC-FS) for the feature selection problem. EMC-FS is a meta-heuristic approach which is quite similar to a simulated annealing algorithm. The difference is that no cooling schedule is required. Improved solutions are accepted and worse solutions are adaptively accepted based on the quality of the trial solution, the search time and the number of consecutive non-improving iterations. We have evaluated our approach against the latest methodologies in the literature on standard benchmark problems. The quality of the obtained subset of features has also been evaluated in terms of the number of generated rules (descriptive patterns) and classification accuracy. Our research demonstrates that our approach produces some of the best known results.",
keywords = "Exponential Monte-Carlo, Feature selection, Local search",
author = "Salwani Abdullah and Sabar, {Nasser R.} and {Ahmad Nazri}, {Mohd Zakree} and Masri Ayob",
year = "2014",
month = "1",
doi = "10.1016/j.cie.2013.10.009",
language = "English",
volume = "67",
pages = "160--167",
journal = "Computers and Industrial Engineering",
issn = "0360-8352",
publisher = "Elsevier Limited",
number = "1",

}

TY - JOUR

T1 - An Exponential Monte-Carlo algorithm for feature selection problems

AU - Abdullah, Salwani

AU - Sabar, Nasser R.

AU - Ahmad Nazri, Mohd Zakree

AU - Ayob, Masri

PY - 2014/1

Y1 - 2014/1

N2 - Feature selection problems (FS) can be defined as the process of eliminating redundant features while avoiding information loss. Due to that fact that FS is an NP-hard problem, heuristic and meta-heuristic approaches have been widely used by researchers. In this work, we proposed an Exponential Monte-Carlo algorithm (EMC-FS) for the feature selection problem. EMC-FS is a meta-heuristic approach which is quite similar to a simulated annealing algorithm. The difference is that no cooling schedule is required. Improved solutions are accepted and worse solutions are adaptively accepted based on the quality of the trial solution, the search time and the number of consecutive non-improving iterations. We have evaluated our approach against the latest methodologies in the literature on standard benchmark problems. The quality of the obtained subset of features has also been evaluated in terms of the number of generated rules (descriptive patterns) and classification accuracy. Our research demonstrates that our approach produces some of the best known results.

AB - Feature selection problems (FS) can be defined as the process of eliminating redundant features while avoiding information loss. Due to that fact that FS is an NP-hard problem, heuristic and meta-heuristic approaches have been widely used by researchers. In this work, we proposed an Exponential Monte-Carlo algorithm (EMC-FS) for the feature selection problem. EMC-FS is a meta-heuristic approach which is quite similar to a simulated annealing algorithm. The difference is that no cooling schedule is required. Improved solutions are accepted and worse solutions are adaptively accepted based on the quality of the trial solution, the search time and the number of consecutive non-improving iterations. We have evaluated our approach against the latest methodologies in the literature on standard benchmark problems. The quality of the obtained subset of features has also been evaluated in terms of the number of generated rules (descriptive patterns) and classification accuracy. Our research demonstrates that our approach produces some of the best known results.

KW - Exponential Monte-Carlo

KW - Feature selection

KW - Local search

UR - http://www.scopus.com/inward/record.url?scp=84889676768&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84889676768&partnerID=8YFLogxK

U2 - 10.1016/j.cie.2013.10.009

DO - 10.1016/j.cie.2013.10.009

M3 - Article

VL - 67

SP - 160

EP - 167

JO - Computers and Industrial Engineering

JF - Computers and Industrial Engineering

SN - 0360-8352

IS - 1

ER -