Hybrid search of feature subsets

Manoranjan Dash, Huan Liu

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

7 Scopus citations

Abstract

Feature selection is a search problem for an "optimal" subset of features. The class separability is normally used as one of the basic feature selection criteria. Instead of maximizing the class separability as in the literature, this work adopts a criterion aiming to maintain the discriminating power of the data. After examining the pros and cons of two existing algorithms for feature selection, we propose a hybrid algorithm of probabilistic and complete search that can take advantage of both algorithms. It begins by rtmning LVF (probabilistic search) to reduce the number of features; then it runs "Automatic Branch & Bound (ABB)" (complete search). By imposing a limit on the amount of time this algorithm can run, we obtain an approximation algorithm. The empirical study suggests that dividing the time equally between the two phases yields nearly the best performance, and that the hybrid search algorithm substantially outperforms earlier methods in general.

Original languageEnglish (US)
Title of host publicationPRICAI 1998
Subtitle of host publicationTopics in Artificial Intelligence - 5th Pacific Rim International Conference on Artificial Intelligence, Proceedings
EditorsHing-Yan Lee, Hiroshi Motoda
PublisherSpringer Verlag
Pages238-249
Number of pages12
ISBN (Print)354065271X, 9783540652717
DOIs
StatePublished - Jan 1 1998
Externally publishedYes
Event5th Pacific Rim Intemational Conference on Artificial Intelligence, PRICAI 1998 - Singapore, Singapore
Duration: Nov 22 1998Nov 27 1998

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume1531
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

Other5th Pacific Rim Intemational Conference on Artificial Intelligence, PRICAI 1998
CountrySingapore
CitySingapore
Period11/22/9811/27/98

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Fingerprint Dive into the research topics of 'Hybrid search of feature subsets'. Together they form a unique fingerprint.

  • Cite this

    Dash, M., & Liu, H. (1998). Hybrid search of feature subsets. In H-Y. Lee, & H. Motoda (Eds.), PRICAI 1998: Topics in Artificial Intelligence - 5th Pacific Rim International Conference on Artificial Intelligence, Proceedings (pp. 238-249). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 1531). Springer Verlag. https://doi.org/10.1007/BFb0095273