The Distance Standard Deviation

Dominic Edelmann; Donald  Richards; Daniel Vogel

doi:10.1214/19-AOS1935

The Distance Standard Deviation

Dominic Edelmann, Donald Richards, Daniel Vogel

Physics

Research output: Contribution to journal › Article › peer-review

7 Citations (Scopus)

Abstract

The distance standard deviation, which arises in distance correlation analysis of multivariate data, is studied as a measure of spread. The asymptotic distribution of the empirical distance standard deviation is derived under the assumption of finite second moments. Applications are provided to hypothesis testing on a data set from materials science and to multivariate statistical quality control. The distance standard deviation is compared to classical scale measures for inference on the spread of heavy-tailed distributions. Inequalities for the distance variance are derived, proving that the distance standard deviation is bounded above by the classical standard deviation and by Gini’s mean difference. New expressions for the distance standard deviation are obtained in terms of Gini’s mean difference and the moments of spacings of order statistics. It is also shown that the distance standard deviation satisfies the axiomatic properties of a measure of spread.

Original language	English
Pages (from-to)	3395-3416
Number of pages	24
Journal	Annals of Statistics
Volume	48
Issue number	6
Early online date	11 Dec 2020
DOIs	https://doi.org/10.1214/19-AOS1935 https://doi.org/10.1214/19-AOS1935SUPP
Publication status	Published - 11 Dec 2020

Bibliographical note

Supplement to “The distance standard deviation”. The supplementary material consists of six appendices. Appendix A contains various proofs to results in the main paper. Appendix B contains details on the derivation of Table 1 in the main document. In Appendix C, a limit theorem for the squared distance covariance is stated, under weaker assumptions than known previously. Simulation results for permutation-based two-sample scale tests are provided in Appendix D. Appendix E gives additional theoretical results for the distance standard deviation in one dimension. Appendix F tabulates the distance variances for a collection of distributions.

Access to Document

10.1214/19-AOS1935Licence: Unspecified
10.1214/19-AOS1935SUPPLicence: Unspecified

Cite this

@article{dc6c8153dea145cdb13dfb1614fbc54c,

title = "The Distance Standard Deviation",

abstract = "The distance standard deviation, which arises in distance correlation analysis of multivariate data, is studied as a measure of spread. The asymptotic distribution of the empirical distance standard deviation is derived under the assumption of finite second moments. Applications are provided to hypothesis testing on a data set from materials science and to multivariate statistical quality control. The distance standard deviation is compared to classical scale measures for inference on the spread of heavy-tailed distributions. Inequalities for the distance variance are derived, proving that the distance standard deviation is bounded above by the classical standard deviation and by Gini{\textquoteright}s mean difference. New expressions for the distance standard deviation are obtained in terms of Gini{\textquoteright}s mean difference and the moments of spacings of order statistics. It is also shown that the distance standard deviation satisfies the axiomatic properties of a measure of spread.",

author = "Dominic Edelmann and Donald Richards and Daniel Vogel",

note = "Supplement to “The distance standard deviation”. The supplementary material consists of six appendices. Appendix A contains various proofs to results in the main paper. Appendix B contains details on the derivation of Table 1 in the main document. In Appendix C, a limit theorem for the squared distance covariance is stated, under weaker assumptions than known previously. Simulation results for permutation-based two-sample scale tests are provided in Appendix D. Appendix E gives additional theoretical results for the distance standard deviation in one dimension. Appendix F tabulates the distance variances for a collection of distributions. ",

year = "2020",

month = dec,

day = "11",

doi = "10.1214/19-AOS1935",

language = "English",

volume = "48",

pages = "3395--3416",

journal = "Annals of Statistics",

issn = "0090-5364",

publisher = "Institute of Mathematical Statistics",

number = "6",

}

TY - JOUR

T1 - The Distance Standard Deviation

AU - Edelmann, Dominic

AU - Richards, Donald

AU - Vogel, Daniel

N1 - Supplement to “The distance standard deviation”. The supplementary material consists of six appendices. Appendix A contains various proofs to results in the main paper. Appendix B contains details on the derivation of Table 1 in the main document. In Appendix C, a limit theorem for the squared distance covariance is stated, under weaker assumptions than known previously. Simulation results for permutation-based two-sample scale tests are provided in Appendix D. Appendix E gives additional theoretical results for the distance standard deviation in one dimension. Appendix F tabulates the distance variances for a collection of distributions.

PY - 2020/12/11

Y1 - 2020/12/11

N2 - The distance standard deviation, which arises in distance correlation analysis of multivariate data, is studied as a measure of spread. The asymptotic distribution of the empirical distance standard deviation is derived under the assumption of finite second moments. Applications are provided to hypothesis testing on a data set from materials science and to multivariate statistical quality control. The distance standard deviation is compared to classical scale measures for inference on the spread of heavy-tailed distributions. Inequalities for the distance variance are derived, proving that the distance standard deviation is bounded above by the classical standard deviation and by Gini’s mean difference. New expressions for the distance standard deviation are obtained in terms of Gini’s mean difference and the moments of spacings of order statistics. It is also shown that the distance standard deviation satisfies the axiomatic properties of a measure of spread.

AB - The distance standard deviation, which arises in distance correlation analysis of multivariate data, is studied as a measure of spread. The asymptotic distribution of the empirical distance standard deviation is derived under the assumption of finite second moments. Applications are provided to hypothesis testing on a data set from materials science and to multivariate statistical quality control. The distance standard deviation is compared to classical scale measures for inference on the spread of heavy-tailed distributions. Inequalities for the distance variance are derived, proving that the distance standard deviation is bounded above by the classical standard deviation and by Gini’s mean difference. New expressions for the distance standard deviation are obtained in terms of Gini’s mean difference and the moments of spacings of order statistics. It is also shown that the distance standard deviation satisfies the axiomatic properties of a measure of spread.

U2 - 10.1214/19-AOS1935

DO - 10.1214/19-AOS1935

M3 - Article

SN - 0090-5364

VL - 48

SP - 3395

EP - 3416

JO - Annals of Statistics

JF - Annals of Statistics

IS - 6

ER -

The Distance Standard Deviation

Abstract

Bibliographical note

Access to Document

Fingerprint

Cite this