Robust relative error estimation

研究成果: ジャーナルへの寄稿記事

1 引用 (Scopus)

抄録

Relative error estimation has been recently used in regression analysis. A crucial issue of the existing relative error estimation procedures is that they are sensitive to outliers. To address this issue, we employ the γ-likelihood function, which is constructed through γ-cross entropy with keeping the original statistical model in use. The estimating equation has a redescending property, a desirable property in robust statistics, for a broad class of noise distributions. To find a minimizer of the negative γ-likelihood function, a majorize-minimization (MM) algorithm is constructed. The proposed algorithm is guaranteed to decrease the negative γ-likelihood function at each iteration. We also derive asymptotic normality of the corresponding estimator together with a simple consistent estimator of the asymptotic covariance matrix, so that we can readily construct approximate confidence sets. Monte Carlo simulation is conducted to investigate the effectiveness of the proposed procedure. Real data analysis illustrates the usefulness of our proposed procedure.

元の言語英語
記事番号632
ジャーナルEntropy
20
発行部数9
DOI
出版物ステータス出版済み - 8 24 2018

Fingerprint

estimators
normality
iteration
regression analysis
confidence
estimating
statistics
entropy
optimization
simulation

All Science Journal Classification (ASJC) codes

  • Physics and Astronomy(all)

これを引用

Robust relative error estimation. / Hirose, Kei; Masuda, Hiroki.

:: Entropy, 巻 20, 番号 9, 632, 24.08.2018.

研究成果: ジャーナルへの寄稿記事

@article{392718d9e4374d9797e71b7ba40b8689,
title = "Robust relative error estimation",
abstract = "Relative error estimation has been recently used in regression analysis. A crucial issue of the existing relative error estimation procedures is that they are sensitive to outliers. To address this issue, we employ the γ-likelihood function, which is constructed through γ-cross entropy with keeping the original statistical model in use. The estimating equation has a redescending property, a desirable property in robust statistics, for a broad class of noise distributions. To find a minimizer of the negative γ-likelihood function, a majorize-minimization (MM) algorithm is constructed. The proposed algorithm is guaranteed to decrease the negative γ-likelihood function at each iteration. We also derive asymptotic normality of the corresponding estimator together with a simple consistent estimator of the asymptotic covariance matrix, so that we can readily construct approximate confidence sets. Monte Carlo simulation is conducted to investigate the effectiveness of the proposed procedure. Real data analysis illustrates the usefulness of our proposed procedure.",
author = "Kei Hirose and Hiroki Masuda",
year = "2018",
month = "8",
day = "24",
doi = "10.3390/e20090632",
language = "English",
volume = "20",
journal = "Entropy",
issn = "1099-4300",
publisher = "Multidisciplinary Digital Publishing Institute (MDPI)",
number = "9",

}

TY - JOUR

T1 - Robust relative error estimation

AU - Hirose, Kei

AU - Masuda, Hiroki

PY - 2018/8/24

Y1 - 2018/8/24

N2 - Relative error estimation has been recently used in regression analysis. A crucial issue of the existing relative error estimation procedures is that they are sensitive to outliers. To address this issue, we employ the γ-likelihood function, which is constructed through γ-cross entropy with keeping the original statistical model in use. The estimating equation has a redescending property, a desirable property in robust statistics, for a broad class of noise distributions. To find a minimizer of the negative γ-likelihood function, a majorize-minimization (MM) algorithm is constructed. The proposed algorithm is guaranteed to decrease the negative γ-likelihood function at each iteration. We also derive asymptotic normality of the corresponding estimator together with a simple consistent estimator of the asymptotic covariance matrix, so that we can readily construct approximate confidence sets. Monte Carlo simulation is conducted to investigate the effectiveness of the proposed procedure. Real data analysis illustrates the usefulness of our proposed procedure.

AB - Relative error estimation has been recently used in regression analysis. A crucial issue of the existing relative error estimation procedures is that they are sensitive to outliers. To address this issue, we employ the γ-likelihood function, which is constructed through γ-cross entropy with keeping the original statistical model in use. The estimating equation has a redescending property, a desirable property in robust statistics, for a broad class of noise distributions. To find a minimizer of the negative γ-likelihood function, a majorize-minimization (MM) algorithm is constructed. The proposed algorithm is guaranteed to decrease the negative γ-likelihood function at each iteration. We also derive asymptotic normality of the corresponding estimator together with a simple consistent estimator of the asymptotic covariance matrix, so that we can readily construct approximate confidence sets. Monte Carlo simulation is conducted to investigate the effectiveness of the proposed procedure. Real data analysis illustrates the usefulness of our proposed procedure.

UR - http://www.scopus.com/inward/record.url?scp=85053672160&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85053672160&partnerID=8YFLogxK

U2 - 10.3390/e20090632

DO - 10.3390/e20090632

M3 - Article

AN - SCOPUS:85053672160

VL - 20

JO - Entropy

JF - Entropy

SN - 1099-4300

IS - 9

M1 - 632

ER -