Principal quantile regression for sufficient dimension reduction with heteroscedasticity

Chong Wang, Seung Jun Shin, Yichao Wu

Research output: Contribution to journalArticle

Abstract

Sufficient dimension reduction (SDR) is a successful tool for reducing data dimensionality without stringent model assumptions. In practice, data often display heteroscedasticity which is of scientific importance in general but frequently overlooked since a primal goal of most existing statistical methods is to identify conditional mean relationship among variables. In this article, we propose a new SDR method called principal quantile regression (PQR) that efficiently tackles heteroscedasticity. PQR can naturally be extended to a nonlinear version via kernel trick. Asymptotic properties are established and an efficient solution path-based algorithm is provided. Numerical examples based on both simulated and real data demonstrate the PQR’s advantageous performance over existing SDR methods. PQR still performs very competitively even for the case without heteroscedasticity.

Original languageEnglish
Pages (from-to)2114-2140
Number of pages27
JournalElectronic Journal of Statistics
Volume12
Issue number2
DOIs
Publication statusPublished - 2018 Jan 1

Fingerprint

Sufficient Dimension Reduction
Heteroscedasticity
Quantile Regression
Reduction Method
Efficient Solution
Statistical method
Asymptotic Properties
Dimensionality
kernel
Numerical Examples
Path
Demonstrate
Dimension reduction
Quantile regression
Model

Keywords

  • Heteroscedasticity
  • Kernel quantile regression
  • Principal quantile regression
  • Sufficient dimension reduction

ASJC Scopus subject areas

  • Statistics and Probability
  • Statistics, Probability and Uncertainty

Cite this

Principal quantile regression for sufficient dimension reduction with heteroscedasticity. / Wang, Chong; Shin, Seung Jun; Wu, Yichao.

In: Electronic Journal of Statistics, Vol. 12, No. 2, 01.01.2018, p. 2114-2140.

Research output: Contribution to journalArticle

@article{0c18386ce3cf43cbadb8ac1d45a790a2,
title = "Principal quantile regression for sufficient dimension reduction with heteroscedasticity",
abstract = "Sufficient dimension reduction (SDR) is a successful tool for reducing data dimensionality without stringent model assumptions. In practice, data often display heteroscedasticity which is of scientific importance in general but frequently overlooked since a primal goal of most existing statistical methods is to identify conditional mean relationship among variables. In this article, we propose a new SDR method called principal quantile regression (PQR) that efficiently tackles heteroscedasticity. PQR can naturally be extended to a nonlinear version via kernel trick. Asymptotic properties are established and an efficient solution path-based algorithm is provided. Numerical examples based on both simulated and real data demonstrate the PQR’s advantageous performance over existing SDR methods. PQR still performs very competitively even for the case without heteroscedasticity.",
keywords = "Heteroscedasticity, Kernel quantile regression, Principal quantile regression, Sufficient dimension reduction",
author = "Chong Wang and Shin, {Seung Jun} and Yichao Wu",
year = "2018",
month = "1",
day = "1",
doi = "10.1214/18-EJS1432",
language = "English",
volume = "12",
pages = "2114--2140",
journal = "Electronic Journal of Statistics",
issn = "1935-7524",
publisher = "Institute of Mathematical Statistics",
number = "2",

}

TY - JOUR

T1 - Principal quantile regression for sufficient dimension reduction with heteroscedasticity

AU - Wang, Chong

AU - Shin, Seung Jun

AU - Wu, Yichao

PY - 2018/1/1

Y1 - 2018/1/1

N2 - Sufficient dimension reduction (SDR) is a successful tool for reducing data dimensionality without stringent model assumptions. In practice, data often display heteroscedasticity which is of scientific importance in general but frequently overlooked since a primal goal of most existing statistical methods is to identify conditional mean relationship among variables. In this article, we propose a new SDR method called principal quantile regression (PQR) that efficiently tackles heteroscedasticity. PQR can naturally be extended to a nonlinear version via kernel trick. Asymptotic properties are established and an efficient solution path-based algorithm is provided. Numerical examples based on both simulated and real data demonstrate the PQR’s advantageous performance over existing SDR methods. PQR still performs very competitively even for the case without heteroscedasticity.

AB - Sufficient dimension reduction (SDR) is a successful tool for reducing data dimensionality without stringent model assumptions. In practice, data often display heteroscedasticity which is of scientific importance in general but frequently overlooked since a primal goal of most existing statistical methods is to identify conditional mean relationship among variables. In this article, we propose a new SDR method called principal quantile regression (PQR) that efficiently tackles heteroscedasticity. PQR can naturally be extended to a nonlinear version via kernel trick. Asymptotic properties are established and an efficient solution path-based algorithm is provided. Numerical examples based on both simulated and real data demonstrate the PQR’s advantageous performance over existing SDR methods. PQR still performs very competitively even for the case without heteroscedasticity.

KW - Heteroscedasticity

KW - Kernel quantile regression

KW - Principal quantile regression

KW - Sufficient dimension reduction

UR - http://www.scopus.com/inward/record.url?scp=85050408521&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85050408521&partnerID=8YFLogxK

U2 - 10.1214/18-EJS1432

DO - 10.1214/18-EJS1432

M3 - Article

AN - SCOPUS:85050408521

VL - 12

SP - 2114

EP - 2140

JO - Electronic Journal of Statistics

JF - Electronic Journal of Statistics

SN - 1935-7524

IS - 2

ER -