Classifiability and Designs for Sampling

For the problem of classifying observations into one of several distributions conditions are obtained for controlling misclassification errors arbitrarily and uniformly in terms of the structure of the underlying distributions. In order to control errors it is pointed out that it is not necessary in...

Full description

Saved in:
Bibliographic Details
Published in:Sankhya. Series A 1974-07, Vol.36 (3), p.237-250
Main Authors: Gupta, Somesh Das, Kinderman, Albert
Format: Article
Language:English
Subjects:
Distance functions
Error rates
Euclidean space
Gaussian distributions
Grants
Mathematical moments
Mathematical theorems
Product distribution
Random variables
Sufficient conditions
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:For the problem of classifying observations into one of several distributions conditions are obtained for controlling misclassification errors arbitrarily and uniformly in terms of the structure of the underlying distributions. In order to control errors it is pointed out that it is not necessary in some cases to draw samples from all the unknown distributions. Classification into multivariate normal distributions is discussed to illustrate the theories.
ISSN:0581-572X