Exploring central subspace via contour regression

Hakbae Lee, Pilkeun Choi

Research output: Contribution to journalArticle

Abstract

Contour regression, a method for estimating the central subspace in regression, is based on estimating contour directions of small variation in the response. These directions span the orthogonal complement of the central subspace and can be extracted according to two measures of variation in the response: simple and general contour regression (SCR and GCR). When the elliptically contoured distribution and mild assumptions hold, the contour regression approach in comparison with existing sufficient dimension reduction methods suggests exhaustiveness of the central space, keeping n-consistency. In addition, the contour-based approach proves robust to violations of departures from ellipticity. In this paper, two kernel simple and general contour regressions (KSCR and KGCR) are proposed and compared with SCR and GCR.

Original languageEnglish
Pages (from-to)9-15
Number of pages7
JournalJournal of the Korean Statistical Society
Volume42
Issue number1
DOIs
Publication statusPublished - 2013 Mar 1

Fingerprint

Central Subspace
Regression
Sufficient Dimension Reduction
Elliptically Contoured Distribution
Ellipticity
Reduction Method
Complement
kernel

All Science Journal Classification (ASJC) codes

  • Statistics and Probability

Cite this

@article{d9268008b8364be592ba619709dd7d1d,
title = "Exploring central subspace via contour regression",
abstract = "Contour regression, a method for estimating the central subspace in regression, is based on estimating contour directions of small variation in the response. These directions span the orthogonal complement of the central subspace and can be extracted according to two measures of variation in the response: simple and general contour regression (SCR and GCR). When the elliptically contoured distribution and mild assumptions hold, the contour regression approach in comparison with existing sufficient dimension reduction methods suggests exhaustiveness of the central space, keeping n-consistency. In addition, the contour-based approach proves robust to violations of departures from ellipticity. In this paper, two kernel simple and general contour regressions (KSCR and KGCR) are proposed and compared with SCR and GCR.",
author = "Hakbae Lee and Pilkeun Choi",
year = "2013",
month = "3",
day = "1",
doi = "10.1016/j.jkss.2012.04.005",
language = "English",
volume = "42",
pages = "9--15",
journal = "Journal of the Korean Statistical Society",
issn = "1226-3192",
publisher = "Korean Statistical Society",
number = "1",

}

Exploring central subspace via contour regression. / Lee, Hakbae; Choi, Pilkeun.

In: Journal of the Korean Statistical Society, Vol. 42, No. 1, 01.03.2013, p. 9-15.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Exploring central subspace via contour regression

AU - Lee, Hakbae

AU - Choi, Pilkeun

PY - 2013/3/1

Y1 - 2013/3/1

N2 - Contour regression, a method for estimating the central subspace in regression, is based on estimating contour directions of small variation in the response. These directions span the orthogonal complement of the central subspace and can be extracted according to two measures of variation in the response: simple and general contour regression (SCR and GCR). When the elliptically contoured distribution and mild assumptions hold, the contour regression approach in comparison with existing sufficient dimension reduction methods suggests exhaustiveness of the central space, keeping n-consistency. In addition, the contour-based approach proves robust to violations of departures from ellipticity. In this paper, two kernel simple and general contour regressions (KSCR and KGCR) are proposed and compared with SCR and GCR.

AB - Contour regression, a method for estimating the central subspace in regression, is based on estimating contour directions of small variation in the response. These directions span the orthogonal complement of the central subspace and can be extracted according to two measures of variation in the response: simple and general contour regression (SCR and GCR). When the elliptically contoured distribution and mild assumptions hold, the contour regression approach in comparison with existing sufficient dimension reduction methods suggests exhaustiveness of the central space, keeping n-consistency. In addition, the contour-based approach proves robust to violations of departures from ellipticity. In this paper, two kernel simple and general contour regressions (KSCR and KGCR) are proposed and compared with SCR and GCR.

UR - http://www.scopus.com/inward/record.url?scp=84872389147&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84872389147&partnerID=8YFLogxK

U2 - 10.1016/j.jkss.2012.04.005

DO - 10.1016/j.jkss.2012.04.005

M3 - Article

AN - SCOPUS:84872389147

VL - 42

SP - 9

EP - 15

JO - Journal of the Korean Statistical Society

JF - Journal of the Korean Statistical Society

SN - 1226-3192

IS - 1

ER -