Transformed estimators of probability density functions with heavy tails and classification

N. M. Markovich

Transformed estimators of probability density functions with heavy tails and classification

N. M. Markovich

Research output: Contribution to journal › Article › peer-review

2 Scopus citations

Abstract

Nonparametric estimation of probability density functions (PDF) with heavy trails is considered. Assumed data are suggested to be transformed to a limited interval. Then PDF are found by inverse transformation of the PDF estimation of the transformed data. The empirical risk of erroneous classification of the empirical Bayesian classifier is proposed as a measure of PDF estimators quality.

Original language	English
Pages (from-to)	118-132
Number of pages	15
Journal	Avtomatika i Telemekhanika
Issue number	4
State	Published - 1 Jan 2002
Externally published	Yes

ASJC Scopus subject areas

Control and Systems Engineering

Cite this

@article{fb132119abae430c91a5608dd36e1da2,

title = "Transformed estimators of probability density functions with heavy tails and classification",

abstract = "Nonparametric estimation of probability density functions (PDF) with heavy trails is considered. Assumed data are suggested to be transformed to a limited interval. Then PDF are found by inverse transformation of the PDF estimation of the transformed data. The empirical risk of erroneous classification of the empirical Bayesian classifier is proposed as a measure of PDF estimators quality.",

author = "Markovich, {N. M.}",

year = "2002",

month = jan,

day = "1",

language = "English",

pages = "118--132",

journal = "Avtomatika i Telemekhanika",

issn = "0005-2310",

publisher = "Nauka, Moscow",

number = "4",

}

TY - JOUR

T1 - Transformed estimators of probability density functions with heavy tails and classification

AU - Markovich, N. M.

PY - 2002/1/1

Y1 - 2002/1/1

N2 - Nonparametric estimation of probability density functions (PDF) with heavy trails is considered. Assumed data are suggested to be transformed to a limited interval. Then PDF are found by inverse transformation of the PDF estimation of the transformed data. The empirical risk of erroneous classification of the empirical Bayesian classifier is proposed as a measure of PDF estimators quality.

AB - Nonparametric estimation of probability density functions (PDF) with heavy trails is considered. Assumed data are suggested to be transformed to a limited interval. Then PDF are found by inverse transformation of the PDF estimation of the transformed data. The empirical risk of erroneous classification of the empirical Bayesian classifier is proposed as a measure of PDF estimators quality.

UR - http://www.scopus.com/inward/record.url?scp=0036411057&partnerID=8YFLogxK

M3 - Article

AN - SCOPUS:0036411057

SN - 0005-2310

SP - 118

EP - 132

JO - Avtomatika i Telemekhanika

JF - Avtomatika i Telemekhanika

IS - 4

ER -

Transformed estimators of probability density functions with heavy tails and classification

Abstract

ASJC Scopus subject areas

Other files and links

Fingerprint

Cite this