Open Access
Translator Disclaimer
December, 1988 Admissibile Kernel Estimators of a Multivariate Density
Daren B. H. Cline
Ann. Statist. 16(4): 1421-1427 (December, 1988). DOI: 10.1214/aos/1176351046

Abstract

A kernel density estimator is defined to be admissible if no other kernel estimator has (among all densities and sample sizes) uniformly smaller mean integrated squared error. Admissible kernel density estimators are precisely those using kernels with nonnegative Fourier transforms bounded by 1. Several examples are given.

Citation

Download Citation

Daren B. H. Cline. "Admissibile Kernel Estimators of a Multivariate Density." Ann. Statist. 16 (4) 1421 - 1427, December, 1988. https://doi.org/10.1214/aos/1176351046

Information

Published: December, 1988
First available in Project Euclid: 12 April 2007

zbMATH: 0653.62027
MathSciNet: MR964931
Digital Object Identifier: 10.1214/aos/1176351046

Subjects:
Primary: 62G05
Secondary: 62C15

Keywords: Admissibility , kernel density estimation

Rights: Copyright © 1988 Institute of Mathematical Statistics

JOURNAL ARTICLE
7 PAGES


SHARE
Vol.16 • No. 4 • December, 1988
Back to Top