Publication detail

The leave-one-out maximum likelihood method for the priestley-chao estimator of conditional density

KONEČNÁ, K.

Original Title

The leave-one-out maximum likelihood method for the priestley-chao estimator of conditional density

Type

conference paper

Language

English

Original Abstract

The contribution is focused on a kernel estimation of conditional density. Kernel smoothing is still popular non-parametric method, in theory as well as in practice. The Priestley-Chao estimator of conditional density is introduced and the statistical properties of the estimator are given. The smoothing parameters called bandwidths play a significant role in kernel smoothing. This is the reason for suggesting the methods for their estimation. The typical approach - the cross-validation method - is supplemented with the leave-one-out maximum log-likelihood method. The performance of the suggested methods is compared via a simulation study and an application on a real data set.

Keywords

kernel smoothing; conditional density; bandwidths; Priestlez-Chao estimator; leave-one-out maximum likelihood method; cross-validation method

Authors

KONEČNÁ, K.

Released

6. 2. 2018

Publisher

Slovak University of Technology in Bratislava in publishing house SPEKTRUM STU

Location

Bratislava

ISBN

978-80-227-4765-3

Book

Proceedings, 17th Conference on Applied Mathematics – APLIMAT 2018

Edition

First edition

Pages from

577

Pages to

589

Pages count

13

URL

http://evlm.stuba.sk/APLIMAT2018/proceedings/

BibTex

@inproceedings{BUT145510,
  author="Kateřina {Pokorová}",
  title="The leave-one-out maximum likelihood method for the priestley-chao estimator of conditional density",
  booktitle="Proceedings, 17th Conference on Applied Mathematics – APLIMAT 2018",
  year="2018",
  series="First edition",
  pages="577--589",
  publisher="Slovak University of Technology in Bratislava in publishing house SPEKTRUM STU",
  address="Bratislava",
  isbn="978-80-227-4765-3",
  url="http://evlm.stuba.sk/APLIMAT2018/proceedings/"
}