CLASSIFICATION OF MULTIPLE OBSERVATIONS USING A RANK NEAREST-NEIGHBOR RULE

被引：0

作者：

BAGUI, SC

机构：

[1] Department of Mathematics and Statistics, The University of West Florida, Pensacola

来源：

PATTERN RECOGNITION LETTERS | 1993年 / 14卷 / 08期

关键词：

BAYES RISK; CLASSIFICATION; MONTE-CARLO; RANK NEAREST NEIGHBOR;

D O I：

10.1016/0167-8655(93)90045-F

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

We consider the problem of classifying multiple (k) univariate observations into one of two populations using a rank nearest neighbor (RNN) type rule. We derive the limiting total probability of misclassification (TPMC) (limiting risk) R(k) of the proposed RNN rule for k = 2 and obtain an upper bound of the limiting TPMC. Finally, with k varying from 1 to 5, some Monte Carlo results are reported to study the performance of the proposed rule.

引用

页码：611 / 617

页数：7