ASYMPTOTIC PROPERTIES OF A KERNEL TYPE ESTIMATOR OF A DENSITY RATIO - Collections

＜journal article＞
ASYMPTOTIC PROPERTIES OF A KERNEL TYPE ESTIMATOR OF A DENSITY RATIO

Creator	Author PID K006885 Creator Name Moriyama, Taku 森山, 卓モリヤマ, タク Affiliation Affiliation Name Graduate School of Mathematics, Kyushu University 九州大学数理学府
Creator	Author PID K002245 Creator Name Maesono, Yoshihiko 前園, 宜彦マエソノ, ヨシヒコ Affiliation Affiliation Name Faculty of Mathematics, Kyushu University 九州大学数理学研究院
Language	English
Publisher	Research Association of Statistical Sciences
Publisher	統計科学研究会
Date	2016-12
Source Title	Bulletin of informatics and cybernetics
Vol	48
First Page	37
Last Page	46
Publication Type	Version of Record
Access Rights	open access
Crossref DOI	https://doi.org/10.5109/2203026
Abstract	In statistical inference, estimation of a density ratio takes important roll. The density ratio is a measure of difference of two populations, and then its estimator is vital. The ratio estimator is u...sed for testing equality of two density functions, detecting change points, discriminant analysis etc. Under some parametric assumptions, there are many papers which study asymptotic properties of the ratio estimators. In this paper we will discuss a kernel type estimator of the ratio.show more

File	FileType	Size	Views	Description
3-Morita	pdf	254 KB	315

PISSN	0286-522X
EISSN	2435-743X
NCID	AA10634475
Record ID	2203026
Subject Terms	Asymptotic properties
	density ratio
	Edgeworth expansion
	kernel type estimator
Funding Information	Funder Name Japan Society for the Promotion of Science 日本学術振興会 Award Number 15K11995 Award Title ノンパラメトリック統計量の平滑化と高精度推測法の構築 Improvement of statistical inference based on nonparametric smoothing statistics
Created Date	2019.02.26
Modified Date	2023.11.21