COMPARISON OF THREE FEATURE SELECTION PROCEDURES FOR SIGNAL ANALYSIS.

Ariel Pascar, Arnon Cohen

Research output: Contribution to conferencePaperpeer-review

Abstract

The feature reduction problem is considered in the context of two equiprobable classes with multivariate pattern vectors. Three methods for feature selection, that are often referred to in the literature, under conditions which deviate from theoretical assumptions, are compared. If the purpose of the feature reduction is well defined and the scatter characteristics of the classes are well estimated, it is possible to determine the optimal method for a specific practical system.

Original languageEnglish
StatePublished - 1 Dec 1985

ASJC Scopus subject areas

  • Engineering (all)

Fingerprint

Dive into the research topics of 'COMPARISON OF THREE FEATURE SELECTION PROCEDURES FOR SIGNAL ANALYSIS.'. Together they form a unique fingerprint.

Cite this