Mutually-exclusive-and-collectively-exhaustive feature selection scheme

Chia-Yen Lee, Bo Syun Chen

研究成果: Article

9 引文 (Scopus)

摘要

In the fields of machine learning and data mining, feature selection methods are used to identify the most cost-effective predictors and to give a deeper understanding of pattern recognition and extraction. This study proposes a novel mutually-exclusive-and-collectively-exhaustive (MECE) feature selection scheme. Based on the MECE principle in decision science, the scheme, which has three stages including evaluation of independence, evaluation of importance and evaluation of completeness, aims to identify the independent and important variables with complete information. A case study of fault classification in semiconductor manufacturing and a study of breast cancer relapse identification in bioinformatics are used to validate the proposed scheme. The results demonstrate that the proposed MECE scheme selects fewer variables, avoids the multicollinearity problem, and improves fault classification accuracy in the two case studies.

原文English
頁(從 - 到)961-971
頁數11
期刊Applied Soft Computing Journal
68
DOIs
出版狀態Published - 2018 七月 1

指紋

Feature extraction
Bioinformatics
Pattern recognition
Data mining
Learning systems
Semiconductor materials
Costs

All Science Journal Classification (ASJC) codes

  • Software

引用此文

@article{4c7f229ec918435c9da3e9c9812ed156,
title = "Mutually-exclusive-and-collectively-exhaustive feature selection scheme",
abstract = "In the fields of machine learning and data mining, feature selection methods are used to identify the most cost-effective predictors and to give a deeper understanding of pattern recognition and extraction. This study proposes a novel mutually-exclusive-and-collectively-exhaustive (MECE) feature selection scheme. Based on the MECE principle in decision science, the scheme, which has three stages including evaluation of independence, evaluation of importance and evaluation of completeness, aims to identify the independent and important variables with complete information. A case study of fault classification in semiconductor manufacturing and a study of breast cancer relapse identification in bioinformatics are used to validate the proposed scheme. The results demonstrate that the proposed MECE scheme selects fewer variables, avoids the multicollinearity problem, and improves fault classification accuracy in the two case studies.",
author = "Chia-Yen Lee and Chen, {Bo Syun}",
year = "2018",
month = "7",
day = "1",
doi = "10.1016/j.asoc.2017.04.055",
language = "English",
volume = "68",
pages = "961--971",
journal = "Applied Soft Computing",
issn = "1568-4946",
publisher = "Elsevier BV",

}

Mutually-exclusive-and-collectively-exhaustive feature selection scheme. / Lee, Chia-Yen; Chen, Bo Syun.

於: Applied Soft Computing Journal, 卷 68, 01.07.2018, p. 961-971.

研究成果: Article

TY - JOUR

T1 - Mutually-exclusive-and-collectively-exhaustive feature selection scheme

AU - Lee, Chia-Yen

AU - Chen, Bo Syun

PY - 2018/7/1

Y1 - 2018/7/1

N2 - In the fields of machine learning and data mining, feature selection methods are used to identify the most cost-effective predictors and to give a deeper understanding of pattern recognition and extraction. This study proposes a novel mutually-exclusive-and-collectively-exhaustive (MECE) feature selection scheme. Based on the MECE principle in decision science, the scheme, which has three stages including evaluation of independence, evaluation of importance and evaluation of completeness, aims to identify the independent and important variables with complete information. A case study of fault classification in semiconductor manufacturing and a study of breast cancer relapse identification in bioinformatics are used to validate the proposed scheme. The results demonstrate that the proposed MECE scheme selects fewer variables, avoids the multicollinearity problem, and improves fault classification accuracy in the two case studies.

AB - In the fields of machine learning and data mining, feature selection methods are used to identify the most cost-effective predictors and to give a deeper understanding of pattern recognition and extraction. This study proposes a novel mutually-exclusive-and-collectively-exhaustive (MECE) feature selection scheme. Based on the MECE principle in decision science, the scheme, which has three stages including evaluation of independence, evaluation of importance and evaluation of completeness, aims to identify the independent and important variables with complete information. A case study of fault classification in semiconductor manufacturing and a study of breast cancer relapse identification in bioinformatics are used to validate the proposed scheme. The results demonstrate that the proposed MECE scheme selects fewer variables, avoids the multicollinearity problem, and improves fault classification accuracy in the two case studies.

UR - http://www.scopus.com/inward/record.url?scp=85019050998&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85019050998&partnerID=8YFLogxK

U2 - 10.1016/j.asoc.2017.04.055

DO - 10.1016/j.asoc.2017.04.055

M3 - Article

AN - SCOPUS:85019050998

VL - 68

SP - 961

EP - 971

JO - Applied Soft Computing

JF - Applied Soft Computing

SN - 1568-4946

ER -