ϵ-Equivalence of Feature Selection Rules

Moshe Ben-Bassat*

*Corresponding author for this work

Research output: Contribution to journalLetterpeer-review

Abstract

The ranking of features according to a feature selection criteria is examined. The concept of ϵ-equivalence is introduced to measure the extent to which a ranking deviates from the ranking induced by the probability of error rule. The relationship between the ϵ-equivalence of a given rule and the bounds on the probability of error derived from this rule is demonstrated. Illustrations of the ϵ-equivalence concept are presented for Shannon's equivocation rule, the quadratic equivocation rule, and the Bhattacharyya rule. A numerical example concludes the presentation.

Original languageEnglish
Pages (from-to)769-772
Number of pages4
JournalIEEE Transactions on Information Theory
Volume24
Issue number6
DOIs
StatePublished - Nov 1978

Fingerprint

Dive into the research topics of 'ϵ-Equivalence of Feature Selection Rules'. Together they form a unique fingerprint.

Cite this