DocumentCode :
923233
Title :
k-nearest-neighbor Bayes-risk estimation
Author :
Fukunaga, Keinosuke ; Hostetler, Larry D.
Volume :
21
Issue :
3
fYear :
1975
fDate :
5/1/1975 12:00:00 AM
Firstpage :
285
Lastpage :
293
Abstract :
Nonparametric estimation of the Bayes risk 
using a 
-nearest-neighbor ( -NN) approach is investigated. Estimates of the conditional Bayes error 
for use in an unclassified test sample approach to estimate are derived using maximum-likelihood estimation techniques. By using the volume information as well as the class representations of the -NN\´s to 
, the mean-squared error of the conditional Bayes error estimate is reduced significantly. Simulations are presented to indicate the performance of the estimates using unclassified testing samples.
using a -nearest-neighbor ( -NN) approach is investigated. Estimates of the conditional Bayes error for use in an unclassified test sample approach to estimate are derived using maximum-likelihood estimation techniques. By using the volume information as well as the class representations of the -NN\´s to , the mean-squared error of the conditional Bayes error estimate is reduced significantly. Simulations are presented to indicate the performance of the estimates using unclassified testing samples.Keywords :
Bayes procedures; Nonparametric estimation; Pattern recognition; Correlators; Delay; Envelope detectors; Maximum likelihood detection; Maximum likelihood estimation; Narrowband; Random variables; Sampling methods; Stochastic resonance; Testing;
fLanguage :
English
Journal_Title :
Information Theory, IEEE Transactions on
Publisher :
ieee
ISSN :
0018-9448
Type :
jour
DOI :
10.1109/TIT.1975.1055373
Filename :
1055373
Link To Document :
