PATTERN-CLASSIFICATION USING AN EFFICIENT KNNR

被引：20

作者：

BELKASIM, SO

SHRIDHAR, M

AHMADI, M

机构：

[1] UNIV WINDSOR,DEPT ELECT ENGN,WINDSOR N9B 3P4,ONTARIO,CANADA

[2] GARYOUNIS UNIV,DEPT ELECT ENGN,BENGHAZI,LIBYA

来源：

PATTERN RECOGNITION | 1992年 / 25卷 / 10期

关键词：

PATTERN CLASSIFICATION; CLASSIFIER; K-NEAREST NEIGHBOR RULE; CHARACTER/SHAPE RECOGNITION; NEIGHBORHOOD CLASSIFIER;

D O I：

10.1016/0031-3203(92)90028-H

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

An efficient method is described to compute the K-nearest neighbour rule (KNNR). The number of distance computations is reduced considerably without any increase in the error rate. Unlike the other techniques, no preprocessing and approximation are involved in this technique which makes it suitable for overlapped classes.

引用

页码：1269 / 1274

页数：6

共 24 条

[1]

ANDERSON TW, 1958, INTRO MULTIVARIATE S, P19

[2]

BELKASIM S, 1990, THESIS U WINDSOR WIN

[3] A FAST CLASSIFIER FOR IMAGE DATA [J].