An Upper Bound on the Asymptotic Error Probability of the k-Nearest Neighbor Rule for Multiple Classes

L. Györfi, Zoltán Györfi

Research output: Contribution to journalArticle

8 Citations (Scopus)

Abstract

If Rk denotes the asymptotic error probability of the k-nearest neighbor rule for m classes and R* denotes the Bayes probability of error, then conditions are given that yield Rk − R* ≼ √MR1/k.

Original languageEnglish
Pages (from-to)512-514
Number of pages3
JournalIEEE Transactions on Information Theory
Volume24
Issue number4
DOIs
Publication statusPublished - 1978

Fingerprint

Error probability

ASJC Scopus subject areas

  • Computer Science Applications
  • Information Systems
  • Library and Information Sciences

Cite this

An Upper Bound on the Asymptotic Error Probability of the k-Nearest Neighbor Rule for Multiple Classes. / Györfi, L.; Györfi, Zoltán.

In: IEEE Transactions on Information Theory, Vol. 24, No. 4, 1978, p. 512-514.

Research output: Contribution to journalArticle

@article{eaa470c2dc954694adcf1da9c42bd409,
title = "An Upper Bound on the Asymptotic Error Probability of the k-Nearest Neighbor Rule for Multiple Classes",
abstract = "If Rk denotes the asymptotic error probability of the k-nearest neighbor rule for m classes and R* denotes the Bayes probability of error, then conditions are given that yield Rk − R* ≼ √MR1/k.",
author = "L. Gy{\"o}rfi and Zolt{\'a}n Gy{\"o}rfi",
year = "1978",
doi = "10.1109/TIT.1978.1055900",
language = "English",
volume = "24",
pages = "512--514",
journal = "IEEE Transactions on Information Theory",
issn = "0018-9448",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "4",

}

TY - JOUR

T1 - An Upper Bound on the Asymptotic Error Probability of the k-Nearest Neighbor Rule for Multiple Classes

AU - Györfi, L.

AU - Györfi, Zoltán

PY - 1978

Y1 - 1978

N2 - If Rk denotes the asymptotic error probability of the k-nearest neighbor rule for m classes and R* denotes the Bayes probability of error, then conditions are given that yield Rk − R* ≼ √MR1/k.

AB - If Rk denotes the asymptotic error probability of the k-nearest neighbor rule for m classes and R* denotes the Bayes probability of error, then conditions are given that yield Rk − R* ≼ √MR1/k.

UR - http://www.scopus.com/inward/record.url?scp=0003961799&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0003961799&partnerID=8YFLogxK

U2 - 10.1109/TIT.1978.1055900

DO - 10.1109/TIT.1978.1055900

M3 - Article

VL - 24

SP - 512

EP - 514

JO - IEEE Transactions on Information Theory

JF - IEEE Transactions on Information Theory

SN - 0018-9448

IS - 4

ER -