What's wrong with Fisher criterion?

被引：44

作者：

Yang, J ^{[1
]}

Yang, JY

Zhang, D

机构：

[1] Nanjing Univ Sci & Technol, Dept Comp Sci, Nanjing 210094, Peoples R China

[2] Hong Kong Polytech Univ, Ctr Multimedia Signal Proc, Dept Comp, Kowloon, Hong Kong, Peoples R China

来源：

PATTERN RECOGNITION | 2002年 / 35卷 / 11期

基金：

中国国家自然科学基金;

关键词：

Fisher criterion; linear discriminant analysis (LDA); feature extraction; handwritten digit recognition;

D O I：

10.1016/S0031-3203(02)00071-7

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

In this paper, a seemingly contradictory experimental result concerning Fisher criterion is first exhibited. Then, we analyze this result from the statistical correlation point of view and give a reasonable explanation. More importantly, we emphasize that Fisher criterion is not an absolute criterion, and, it should be associated with the statistical correlation together to assess the discrimination of a set of discriminant vectors. (C) 2002 Pattern Recognition Society. Published by Elsevier Science Ltd. All rights reserved.

引用

页码：2665 / 2668

页数：4

共 5 条

[1] OPTIMAL SET OF DISCRIMINANT VECTORS [J].

FOLEY, DH ;

SAMMON, JW .

IEEE TRANSACTIONS ON COMPUTERS, 1975, C 24 (03) :281-289

[2]

Hart P.E., 1973, Pattern recognition and scene analysis

[3] A theorem on the uncorrelated optimal discriminant vectors [J].

Jin, Z ;

Yang, JY ;

Tang, ZM ;

Hu, ZS .

PATTERN RECOGNITION, 2001, 34 (10) :2041-2047

[4] Face recognition based on the uncorrelated discriminant transformation [J].

Jin, Z ;

Yang, JY ;

Hu, ZS ;

Lou, Z .

PATTERN RECOGNITION, 2001, 34 (07) :1405-1416

[5] An optimal K-L transform method for feature extraction [J].

Yang, J ;

Yang, JY .

IMAGE EXTRACTION, SEGMENTATION, AND RECOGNITION, 2001, 4550 :239-244

← 1 →