Consistency based feature selection

Manoranjan Dash, Huan Liu, Hiroshi Motoda

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

135 Scopus citations


Feature selection is an effective technique in dealing with dimensionality reduction for classification task, a main component of data mining. It searches for ein "optimal" subset of features. The search strategies under consideration are one of the three: complete, heuristic, and probabilistic. Existing algorithms adopt various measiires to evaluate the goodness of feature subsets. This work focuses on one measure called consistency. We study its properties in compsirison with other major measures and different ways of using this measure in sejirch of feature subsets. We conduct cin empirical study to examine the pros and cons of these different search methods using consistency. Through this extensive exercise, we ciim to provide a comprehensive view of this measure and its relations with other measures cind a guideline of the use of this meeisure with different search strategies facing a new application.

Original languageEnglish (US)
Title of host publicationKnowledge Discovery and Data Mining
Subtitle of host publicationCurrent Issues and New Applications - 4th Pacific-Asia Conference, PAKDD 2000, Proceedings
EditorsTakao Terano, Huan Liu, Arbee L.P. Chen
PublisherSpringer Verlag
Number of pages12
ISBN (Print)3540673822, 9783540673828
StatePublished - 2000
Externally publishedYes
Event4th Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2000 - Kyoto, Japan
Duration: Apr 18 2000Apr 20 2000

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349


Other4th Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2000

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)


Dive into the research topics of 'Consistency based feature selection'. Together they form a unique fingerprint.

Cite this