Nonparametric multivariate kurtosis and tailweight measures

Jin Wang, Robert Serfling

Research output: Contribution to journalArticle

15 Citations (Scopus)

Abstract

For nonparametric exploration or description of a distribution, the treatment of location, spread, symmetry, and skewness is followed by characterization of kurtosis. Classical moment-based kurtosis measures the dispersion of a distribution about its 'shoulders'. Here, we consider quantile-based kurtosis measures. These are robust, defined more widely, and discriminate better among shapes. A univariate quantile-based kurtosis measure of [Groeneveld, R.A. and Meeden, G., 1984, Measuring skewness and kurtosis. The Statistician, 33, 391-399.] is extended to the multivariate case by representing it as a transform of a dispersion functional. A family of such kurtosis measures defined for a given distribution and taken together comprises a real-valued 'kurtosis functional', which has intuitive appeal as a convenient two-dimensional curve for description of the kurtosis of the distribution. Several multivariate distributions in any dimension may thus be compared with respect to their kurtosis in a single two-dimensional plot. Important properties of the new multivariate kurtosis measures are established. For example, for elliptically symmetric distributions, this measure determines the distribution within affine equivalence. Related tailweight measures, influence curves, and asymptotic behavior of sample versions are also discussed.

Original languageEnglish (US)
Pages (from-to)441-456
Number of pages16
JournalJournal of Nonparametric Statistics
Volume17
Issue number4
DOIs
StatePublished - Jun 2005

Fingerprint

Kurtosis
Skewness
Quantile
Elliptically Symmetric Distributions
Influence Curve
Appeal
Multivariate Distribution
Univariate
Intuitive
Asymptotic Behavior
Equivalence
Transform
Moment
Symmetry
Curve

Keywords

  • Depth functions
  • Influence curves
  • Kurtosis
  • Tailweight

ASJC Scopus subject areas

  • Mathematics(all)
  • Statistics and Probability

Cite this

Nonparametric multivariate kurtosis and tailweight measures. / Wang, Jin; Serfling, Robert.

In: Journal of Nonparametric Statistics, Vol. 17, No. 4, 06.2005, p. 441-456.

Research output: Contribution to journalArticle

@article{9540f3eb79024362a1ca984ba3178ed2,
title = "Nonparametric multivariate kurtosis and tailweight measures",
abstract = "For nonparametric exploration or description of a distribution, the treatment of location, spread, symmetry, and skewness is followed by characterization of kurtosis. Classical moment-based kurtosis measures the dispersion of a distribution about its 'shoulders'. Here, we consider quantile-based kurtosis measures. These are robust, defined more widely, and discriminate better among shapes. A univariate quantile-based kurtosis measure of [Groeneveld, R.A. and Meeden, G., 1984, Measuring skewness and kurtosis. The Statistician, 33, 391-399.] is extended to the multivariate case by representing it as a transform of a dispersion functional. A family of such kurtosis measures defined for a given distribution and taken together comprises a real-valued 'kurtosis functional', which has intuitive appeal as a convenient two-dimensional curve for description of the kurtosis of the distribution. Several multivariate distributions in any dimension may thus be compared with respect to their kurtosis in a single two-dimensional plot. Important properties of the new multivariate kurtosis measures are established. For example, for elliptically symmetric distributions, this measure determines the distribution within affine equivalence. Related tailweight measures, influence curves, and asymptotic behavior of sample versions are also discussed.",
keywords = "Depth functions, Influence curves, Kurtosis, Tailweight",
author = "Jin Wang and Robert Serfling",
year = "2005",
month = "6",
doi = "10.1080/10485250500039130",
language = "English (US)",
volume = "17",
pages = "441--456",
journal = "Journal of Nonparametric Statistics",
issn = "1048-5252",
publisher = "Taylor and Francis Ltd.",
number = "4",

}

TY - JOUR

T1 - Nonparametric multivariate kurtosis and tailweight measures

AU - Wang, Jin

AU - Serfling, Robert

PY - 2005/6

Y1 - 2005/6

N2 - For nonparametric exploration or description of a distribution, the treatment of location, spread, symmetry, and skewness is followed by characterization of kurtosis. Classical moment-based kurtosis measures the dispersion of a distribution about its 'shoulders'. Here, we consider quantile-based kurtosis measures. These are robust, defined more widely, and discriminate better among shapes. A univariate quantile-based kurtosis measure of [Groeneveld, R.A. and Meeden, G., 1984, Measuring skewness and kurtosis. The Statistician, 33, 391-399.] is extended to the multivariate case by representing it as a transform of a dispersion functional. A family of such kurtosis measures defined for a given distribution and taken together comprises a real-valued 'kurtosis functional', which has intuitive appeal as a convenient two-dimensional curve for description of the kurtosis of the distribution. Several multivariate distributions in any dimension may thus be compared with respect to their kurtosis in a single two-dimensional plot. Important properties of the new multivariate kurtosis measures are established. For example, for elliptically symmetric distributions, this measure determines the distribution within affine equivalence. Related tailweight measures, influence curves, and asymptotic behavior of sample versions are also discussed.

AB - For nonparametric exploration or description of a distribution, the treatment of location, spread, symmetry, and skewness is followed by characterization of kurtosis. Classical moment-based kurtosis measures the dispersion of a distribution about its 'shoulders'. Here, we consider quantile-based kurtosis measures. These are robust, defined more widely, and discriminate better among shapes. A univariate quantile-based kurtosis measure of [Groeneveld, R.A. and Meeden, G., 1984, Measuring skewness and kurtosis. The Statistician, 33, 391-399.] is extended to the multivariate case by representing it as a transform of a dispersion functional. A family of such kurtosis measures defined for a given distribution and taken together comprises a real-valued 'kurtosis functional', which has intuitive appeal as a convenient two-dimensional curve for description of the kurtosis of the distribution. Several multivariate distributions in any dimension may thus be compared with respect to their kurtosis in a single two-dimensional plot. Important properties of the new multivariate kurtosis measures are established. For example, for elliptically symmetric distributions, this measure determines the distribution within affine equivalence. Related tailweight measures, influence curves, and asymptotic behavior of sample versions are also discussed.

KW - Depth functions

KW - Influence curves

KW - Kurtosis

KW - Tailweight

UR - http://www.scopus.com/inward/record.url?scp=16644394582&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=16644394582&partnerID=8YFLogxK

U2 - 10.1080/10485250500039130

DO - 10.1080/10485250500039130

M3 - Article

AN - SCOPUS:16644394582

VL - 17

SP - 441

EP - 456

JO - Journal of Nonparametric Statistics

JF - Journal of Nonparametric Statistics

SN - 1048-5252

IS - 4

ER -