Nonparametric inference for stochastic linear hypotheses

Application to high-dimensional data

Jeanne Kowalski, Jonathan Powell

Research output: Contribution to journalArticle

Abstract

The Mann-Whitney-Wilcoxon rank sum test is limited to comparison of two groups with univariate responses. In this paper, we introduce a class of stochastic linear hypotheses that addresses these limitations within a nonparametric setting. We formulate hypotheses for simultaneous comparisons of several, multivariate response groups, without modelling the response distributions. Inference is developed based on U-statistics theory and an exchangeability assumption. The latter condition is required to identify testable hypotheses for high-dimensional response vectors, such as those arising in genomic and psychosocial research. The methodology is illustrated with two real-data applications.

Original languageEnglish (US)
Pages (from-to)393-408
Number of pages16
JournalBiometrika
Volume91
Issue number2
DOIs
StatePublished - 2004

Fingerprint

Linear Hypothesis
Nonparametric Inference
High-dimensional Data
Nonparametric Statistics
statistics
Statistics
genomics
Wilcoxon rank-sum test
Multivariate Response
Exchangeability
U-statistics
testing
Univariate
Genomics
High-dimensional
methodology
Research
Methodology
Modeling
Inference

Keywords

  • Asymptotic distribution
  • Exchangeability assumption
  • U-statistic

ASJC Scopus subject areas

  • Agricultural and Biological Sciences(all)
  • Agricultural and Biological Sciences (miscellaneous)
  • Statistics and Probability
  • Mathematics(all)
  • Applied Mathematics

Cite this

Nonparametric inference for stochastic linear hypotheses : Application to high-dimensional data. / Kowalski, Jeanne; Powell, Jonathan.

In: Biometrika, Vol. 91, No. 2, 2004, p. 393-408.

Research output: Contribution to journalArticle

@article{83f8143de13843668ad90c3a6cfad528,
title = "Nonparametric inference for stochastic linear hypotheses: Application to high-dimensional data",
abstract = "The Mann-Whitney-Wilcoxon rank sum test is limited to comparison of two groups with univariate responses. In this paper, we introduce a class of stochastic linear hypotheses that addresses these limitations within a nonparametric setting. We formulate hypotheses for simultaneous comparisons of several, multivariate response groups, without modelling the response distributions. Inference is developed based on U-statistics theory and an exchangeability assumption. The latter condition is required to identify testable hypotheses for high-dimensional response vectors, such as those arising in genomic and psychosocial research. The methodology is illustrated with two real-data applications.",
keywords = "Asymptotic distribution, Exchangeability assumption, U-statistic",
author = "Jeanne Kowalski and Jonathan Powell",
year = "2004",
doi = "10.1093/biomet/91.2.393",
language = "English (US)",
volume = "91",
pages = "393--408",
journal = "Biometrika",
issn = "0006-3444",
publisher = "Oxford University Press",
number = "2",

}

TY - JOUR

T1 - Nonparametric inference for stochastic linear hypotheses

T2 - Application to high-dimensional data

AU - Kowalski, Jeanne

AU - Powell, Jonathan

PY - 2004

Y1 - 2004

N2 - The Mann-Whitney-Wilcoxon rank sum test is limited to comparison of two groups with univariate responses. In this paper, we introduce a class of stochastic linear hypotheses that addresses these limitations within a nonparametric setting. We formulate hypotheses for simultaneous comparisons of several, multivariate response groups, without modelling the response distributions. Inference is developed based on U-statistics theory and an exchangeability assumption. The latter condition is required to identify testable hypotheses for high-dimensional response vectors, such as those arising in genomic and psychosocial research. The methodology is illustrated with two real-data applications.

AB - The Mann-Whitney-Wilcoxon rank sum test is limited to comparison of two groups with univariate responses. In this paper, we introduce a class of stochastic linear hypotheses that addresses these limitations within a nonparametric setting. We formulate hypotheses for simultaneous comparisons of several, multivariate response groups, without modelling the response distributions. Inference is developed based on U-statistics theory and an exchangeability assumption. The latter condition is required to identify testable hypotheses for high-dimensional response vectors, such as those arising in genomic and psychosocial research. The methodology is illustrated with two real-data applications.

KW - Asymptotic distribution

KW - Exchangeability assumption

KW - U-statistic

UR - http://www.scopus.com/inward/record.url?scp=3042853652&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=3042853652&partnerID=8YFLogxK

U2 - 10.1093/biomet/91.2.393

DO - 10.1093/biomet/91.2.393

M3 - Article

VL - 91

SP - 393

EP - 408

JO - Biometrika

JF - Biometrika

SN - 0006-3444

IS - 2

ER -