Groupwise Dimension Reduction via Envelope Method

Zifang Guo, Lexin Li, Wenbin Lu, Bing Li

Research output: Contribution to journalArticle

9 Citations (Scopus)

Abstract

The family of sufficient dimension reduction (SDR) methods that produce informative combinations of predictors, or indices, are particularly useful for high-dimensional regression analysis. In many such analyses, it becomes increasingly common that there is available a priori subject knowledge of the predictors; for example, they belong to different groups. While many recent SDR proposals have greatly expanded the scope of the methods’ applicability, how to effectively incorporate the prior predictor structure information remains a challenge. In this article, we aim at dimension reduction that recovers full regression information while preserving the predictor group structure. Built upon a new concept of the direct sum envelope, we introduce a systematic way to incorporate the group information in most existing SDR estimators. As a result, the reduction outcomes are much easier to interpret. Moreover, the envelope method provides a principled way to build a variety of prior structures into dimension reduction analysis. Both simulations and real data analysis demonstrate the competent numerical performance of the new method.

Original languageEnglish (US)
Pages (from-to)1515-1527
Number of pages13
JournalJournal of the American Statistical Association
Volume110
Issue number512
DOIs
StatePublished - Oct 2 2015

Fingerprint

Sufficient Dimension Reduction
Dimension Reduction
Envelope
Predictors
Information Structure
Dimensional Analysis
Reduction Method
Direct Sum
Regression Analysis
Data analysis
High-dimensional
Regression
Estimator
Dimension reduction
Demonstrate
Simulation

All Science Journal Classification (ASJC) codes

  • Statistics and Probability
  • Statistics, Probability and Uncertainty

Cite this

Guo, Zifang ; Li, Lexin ; Lu, Wenbin ; Li, Bing. / Groupwise Dimension Reduction via Envelope Method. In: Journal of the American Statistical Association. 2015 ; Vol. 110, No. 512. pp. 1515-1527.
@article{131d73e418cf4e2f938ed3a7cc33532d,
title = "Groupwise Dimension Reduction via Envelope Method",
abstract = "The family of sufficient dimension reduction (SDR) methods that produce informative combinations of predictors, or indices, are particularly useful for high-dimensional regression analysis. In many such analyses, it becomes increasingly common that there is available a priori subject knowledge of the predictors; for example, they belong to different groups. While many recent SDR proposals have greatly expanded the scope of the methods’ applicability, how to effectively incorporate the prior predictor structure information remains a challenge. In this article, we aim at dimension reduction that recovers full regression information while preserving the predictor group structure. Built upon a new concept of the direct sum envelope, we introduce a systematic way to incorporate the group information in most existing SDR estimators. As a result, the reduction outcomes are much easier to interpret. Moreover, the envelope method provides a principled way to build a variety of prior structures into dimension reduction analysis. Both simulations and real data analysis demonstrate the competent numerical performance of the new method.",
author = "Zifang Guo and Lexin Li and Wenbin Lu and Bing Li",
year = "2015",
month = "10",
day = "2",
doi = "10.1080/01621459.2014.970687",
language = "English (US)",
volume = "110",
pages = "1515--1527",
journal = "Journal of the American Statistical Association",
issn = "0162-1459",
publisher = "Taylor and Francis Ltd.",
number = "512",

}

Groupwise Dimension Reduction via Envelope Method. / Guo, Zifang; Li, Lexin; Lu, Wenbin; Li, Bing.

In: Journal of the American Statistical Association, Vol. 110, No. 512, 02.10.2015, p. 1515-1527.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Groupwise Dimension Reduction via Envelope Method

AU - Guo, Zifang

AU - Li, Lexin

AU - Lu, Wenbin

AU - Li, Bing

PY - 2015/10/2

Y1 - 2015/10/2

N2 - The family of sufficient dimension reduction (SDR) methods that produce informative combinations of predictors, or indices, are particularly useful for high-dimensional regression analysis. In many such analyses, it becomes increasingly common that there is available a priori subject knowledge of the predictors; for example, they belong to different groups. While many recent SDR proposals have greatly expanded the scope of the methods’ applicability, how to effectively incorporate the prior predictor structure information remains a challenge. In this article, we aim at dimension reduction that recovers full regression information while preserving the predictor group structure. Built upon a new concept of the direct sum envelope, we introduce a systematic way to incorporate the group information in most existing SDR estimators. As a result, the reduction outcomes are much easier to interpret. Moreover, the envelope method provides a principled way to build a variety of prior structures into dimension reduction analysis. Both simulations and real data analysis demonstrate the competent numerical performance of the new method.

AB - The family of sufficient dimension reduction (SDR) methods that produce informative combinations of predictors, or indices, are particularly useful for high-dimensional regression analysis. In many such analyses, it becomes increasingly common that there is available a priori subject knowledge of the predictors; for example, they belong to different groups. While many recent SDR proposals have greatly expanded the scope of the methods’ applicability, how to effectively incorporate the prior predictor structure information remains a challenge. In this article, we aim at dimension reduction that recovers full regression information while preserving the predictor group structure. Built upon a new concept of the direct sum envelope, we introduce a systematic way to incorporate the group information in most existing SDR estimators. As a result, the reduction outcomes are much easier to interpret. Moreover, the envelope method provides a principled way to build a variety of prior structures into dimension reduction analysis. Both simulations and real data analysis demonstrate the competent numerical performance of the new method.

UR - http://www.scopus.com/inward/record.url?scp=84954438362&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84954438362&partnerID=8YFLogxK

U2 - 10.1080/01621459.2014.970687

DO - 10.1080/01621459.2014.970687

M3 - Article

AN - SCOPUS:84954438362

VL - 110

SP - 1515

EP - 1527

JO - Journal of the American Statistical Association

JF - Journal of the American Statistical Association

SN - 0162-1459

IS - 512

ER -