Novel algorithms of attribute reduction for variable precision rough set

Yan-Yan YANG, De-Gang CHEN, Sam KWONG

Research output: Book Chapters | Papers in Conference ProceedingsConference paper (refereed)Researchpeer-review

1 Citation (Scopus)

Abstract

The main application of variable precision rough set is to perform attribute reduction for databases. In variable precision rough set, the approach of discernibility matrix is theoretical foundation of finding reducts. In this paper, we observe that only minimal elements in the discernibility matrix is sufficient to find reducts, and every minimal element in the discernibility matrix is determined by one equivalence class pair relative to condition attributes at least; this fact motivates our idea in this paper to search the connection between this kind of pair and the minimal element in the discernibility matrix. By the connection between them, we develop the novel algorithms of finding reducts, which improve the existing ones in terms of discernibility matrix. © 2011 IEEE.
Original languageEnglish
Title of host publicationProceedings of 2011 International Conference on Machine Learning and Cybernetics
PublisherIEEE
Pages108-112
Number of pages5
ISBN (Electronic)9781457703089
ISBN (Print)9781457703058
DOIs
Publication statusPublished - Jul 2011
Externally publishedYes
Event2011 International Conference on Machine Learning and Cybernetics, ICMLC 2011 - Guilin, Guangxi, China
Duration: 10 Jul 201113 Jul 2011

Conference

Conference2011 International Conference on Machine Learning and Cybernetics, ICMLC 2011
Country/TerritoryChina
CityGuilin, Guangxi
Period10/07/1113/07/11

Keywords

  • Discernibility matrix
  • Equivalence class pair relative to condition attributes
  • Minimal element
  • Variable precision rough set

Fingerprint

Dive into the research topics of 'Novel algorithms of attribute reduction for variable precision rough set'. Together they form a unique fingerprint.

Cite this