MINIMUM ENTROPY OF ERROR ESTIMATION FOR DISCRETE RANDOM-VARIABLES

Authors
JANZURA M KOSKI T OTAHAL A
Citation
M. Janzura et al., MINIMUM ENTROPY OF ERROR ESTIMATION FOR DISCRETE RANDOM-VARIABLES, IEEE transactions on information theory, 42(4), 1996, pp. 1193-1201
Citations number
25
Categorie Soggetti
Information Science & Library Science","Engineering, Eletrical & Electronic
Journal title
IEEE transactions on information theory → ACNP
ISSN journal
00189448
Volume
42
Issue
4
Year of publication
1996
Pages
1193 - 1201
Database
ISI
SICI code
0018-9448(1996)42:4<1193:MEOEEF>2.0.ZU;2-9
Abstract
The principle of minimum entropy of error estimation (MEEE) is formula ted for discrete random variables. In the case when the random variabl e to be estimated is binary, me show that the MEEE is given by a Neyma n-Pearson-type strictly monotonous test, In addition, the asymptotic b ehavior of the error probabilities is proved to be equivalent to that of the Bayesian test.