FLOATING SEARCH METHODS IN FEATURE-SELECTION

被引：2128

作者：

PUDIL, P ^{[1
]}

NOVOVICOVA, J ^{[1
]}

KITTLER, J ^{[1
]}

机构：

[1] ACAD SCI CZECH REPUBL,INST INFORMAT THEORY & AUTOMAT,CR-18208 PRAGUE 8,CZECH REPUBLIC

来源：

PATTERN RECOGNITION LETTERS | 1994年 / 15卷 / 11期

关键词：

PATTERN RECOGNITION; FEATURE SELECTION; FEATURE ORDERING; SEARCH METHODS;

D O I：

10.1016/0167-8655(94)90127-9

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

Sequential search methods characterized by a dynamically changing number of features included or eliminated at each step, henceforth ''floating'' methods, are presented. They are shown to give very good results and to be computationally more effective than the branch and bound method.

引用

页码：1119 / 1125

页数：7

共 16 条

[1]

BACKER E, 1977, SEMINAR PATTERN RECO

[2]

CHOAKJARERNWANI.N, 1991, VSSPTR591 U SURR TEC

[3]

CHOAKJARERNWANI.N, 1991, VSSPTR191 U SURR TEC

[4] OPTIMUM FEATURE-SELECTION BY ZERO-ONE INTEGER PROGRAMMING [J].

ICHINO, M ;

SKLANSKY, J .

IEEE TRANSACTIONS ON SYSTEMS MAN AND CYBERNETICS, 1984, 14 (05) :737-746

[5]

Kittler J., 1978, Pattern Recognition and Signal Processing, P41

[6] ON EFFECTIVENESS OF RECEPTORS IN RECOGNITION SYSTEMS [J].

MARILL, T ;

GREEN, DM .

IEEE TRANSACTIONS ON INFORMATION THEORY, 1963, 9 (01) :11-&

[7] EXPERIMENTAL STUDY OF INFORMATION MEASURE AND INTER-INTRA CLASS DISTANCE RATIOS ON FEATURE SELECTION AND ORDERINGS [J].

MICHAEL, M ;

LIN, WC .

IEEE TRANSACTIONS ON SYSTEMS MAN AND CYBERNETICS, 1973, SMC3 (02) :172-181

[8]

NARENDRA P, 1977, IEEE T COMPUT, V26, P917, DOI 10.1109/TC.1977.1674939

[9] AN ANALYSIS OF THE MAX-MIN APPROACH TO FEATURE-SELECTION AND ORDERING [J].

PUDIL, P ;

NOVOVICOVA, J ;

CHOAKJARERNWANIT, N ;

KITTLER, J .

PATTERN RECOGNITION LETTERS, 1993, 14 (11) :841-847

[10]

PUDIL P, 1991, KYBERNETIKA S, V27, P1

← 1 2 →