LINEAR AND QUADRATIC DISCRIMINATION IN PATTERN RECOGNITION

被引：3

作者：

DAY, NE

机构：

[1] Dept. of Statistics, Inst. of Advanced Studies, Australian National University, Canberra, Australia

来源：

IEEE TRANSACTIONS ON INFORMATION THEORY | 1969年 / 15卷 / 03期

关键词：

D O I：

10.1109/TIT.1969.1054315

中图分类号：

TP [自动化技术、计算机技术];

学科分类号：

0812 ;

摘要：

The class of families of density functions for which linear or quadratic discrimination is optimum is obtained. For linear discrimination, this class is the class of linear exponential families of distributions, for quadratic discrimination it is the class of quadratic exponential families. © 1969 IEEE. All rights reserved.

引用

页码：419 / +

共 7 条

[1]

COOPER PW, 1965, IEEE T INFORMATION T, VIT11, P313

[2]

COOPER PW, 1962, CYBERNETICA, V5, P215

[3] A GENERAL MAXIMUM LIKELIHOOD DISCRIMINANT [J].

DAY, NE ;

KERRIDGE, DF .

BIOMETRICS, 1967, 23 (02) :313-&

[4] A SOLUTION TO THE PROBLEM OF OPTIMUM CLASSIFICATION [J].

HOEL, PG ;

PETERSON, RP .

ANNALS OF MATHEMATICAL STATISTICS, 1949, 20 (03) :433-438

[5]

LEHMANN EL, 1959, TESTING STATISTICAL

[6]

SOMMERVILLE DMY, 1934, ANALYTICAL GEOMETRY

[7]

WANI JK, 1968 P CAMBR PHIL 2, V64, P481

← 1 →