A multiclass, k-NN approach to Bayes risk estimation

Authors:
Pierre A Devijver
Affiliations:
Philips Research Laboratory, Ave. Em. Van Becelaere 2, Box 8, Brussels, Belgium
Venue:
Pattern Recognition Letters
Year:
1985

Citing 11
Cited 0

On a New Class of Bounds on Bayes Risk in Multihypothesis Pattern Recognition

IEEE Transactions on Computers
Nonparametric estimation of the bayes error of feature extractors using ordered nearest neighbor sets

IEEE Transactions on Computers
The Probability Distribution of Conditional Classification Error

IEEE Transactions on Pattern Analysis and Machine Intelligence
Statistical Properties of Error Estimators in Performance Assessment of Recognition Systems

IEEE Transactions on Pattern Analysis and Machine Intelligence
Nearest neighbor pattern classification

IEEE Transactions on Information Theory
On optimum recognition error and reject tradeoff

IEEE Transactions on Information Theory
Application of optimum error-reject functions (Corresp.)

IEEE Transactions on Information Theory
Nonparametric Bayes error estimation using unclassified samples

IEEE Transactions on Information Theory
k-nearest-neighbor Bayes-risk estimation

IEEE Transactions on Information Theory
An upper bound on the asymptotic error probability on the k-nearest neighbor rule for multiple classes (Corresp.)

IEEE Transactions on Information Theory
New error bounds with the nearest neighbor rule (Corresp.)

IEEE Transactions on Information Theory

Quantified Score

Hi-index	0.10

Visualization

Abstract

In this paper, the k-NN approach is used for the purpose of estimating the multiclass, 1-NN Bayes error bounds. We derive an estimator which is asymptotically unbiased, and whose variance can be controlled by the choice of k. The estimator appears to be very economic in its use of samples, and quite stable even in very small sample cases.