Title :
On divergence and probability of error in pattern recognition
Author_Institution :
University of California, Irvine, Calif.
fDate :
6/1/1973 12:00:00 AM
Abstract :
An upperbound on the probability of error in classifying a pattern using Bayesian decision criterion is obtained in terms of its divergence and it is shown that the maximization of the divergence of the pattern minimizes this upperbound. Furthermore, a relationship between the divergence of a pattern and its nearest neighbor classification risk is presented.
Keywords :
Bayesian methods; Gaussian processes; Linear matrix inequalities; Marine vehicles; Monte Carlo methods; Nearest neighbor searches; Pattern recognition; Probability density function;
Journal_Title :
Proceedings of the IEEE
DOI :
10.1109/PROC.1973.9164
