A canonical form of Anderson's classification statistic

Yu Shyr

doi:10.1080/03610929808832114

A canonical form of Anderson's classification statistic

Yu Shyr

Department of Statistics

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

Abstract

A canonical reduction of Anderson's Classification Statistic in the case of two multivariate normal populations with different means and the same covariance matrix is obtained in this paper. The statistic and a similar statistic that arise in the estimation of relative potency of a new drug, using a composite response, are expressed in terms of χ², non-central χ^2- and normal variables. The moments of Anderson's statistic are derived and used for obtaining an asymptotic normal approximation to the chance of misclassification.

Original language	English
Pages (from-to)	577-587
Number of pages	11
Journal	Communications in Statistics - Theory and Methods
Volume	27
Issue number	3
DOIs	https://doi.org/10.1080/03610929808832114
Publication status	Published - 1998 Jan 1

All Science Journal Classification (ASJC) codes

Statistics and Probability

Access to Document

10.1080/03610929808832114

Fingerprint Dive into the research topics of 'A canonical form of Anderson's classification statistic'. Together they form a unique fingerprint.

Cite this

@article{e77d7067a9174773a0401428cb9053f7,

title = "A canonical form of Anderson's classification statistic",

abstract = "A canonical reduction of Anderson's Classification Statistic in the case of two multivariate normal populations with different means and the same covariance matrix is obtained in this paper. The statistic and a similar statistic that arise in the estimation of relative potency of a new drug, using a composite response, are expressed in terms of χ2, non-central χ2- and normal variables. The moments of Anderson's statistic are derived and used for obtaining an asymptotic normal approximation to the chance of misclassification.",

author = "Yu Shyr",

year = "1998",

month = jan,

day = "1",

doi = "10.1080/03610929808832114",

language = "English",

volume = "27",

pages = "577--587",

journal = "Communications in Statistics - Theory and Methods",

issn = "0361-0926",

publisher = "Taylor and Francis Ltd.",

number = "3",

}

TY - JOUR

T1 - A canonical form of Anderson's classification statistic

AU - Shyr, Yu

PY - 1998/1/1

Y1 - 1998/1/1

N2 - A canonical reduction of Anderson's Classification Statistic in the case of two multivariate normal populations with different means and the same covariance matrix is obtained in this paper. The statistic and a similar statistic that arise in the estimation of relative potency of a new drug, using a composite response, are expressed in terms of χ2, non-central χ2- and normal variables. The moments of Anderson's statistic are derived and used for obtaining an asymptotic normal approximation to the chance of misclassification.

AB - A canonical reduction of Anderson's Classification Statistic in the case of two multivariate normal populations with different means and the same covariance matrix is obtained in this paper. The statistic and a similar statistic that arise in the estimation of relative potency of a new drug, using a composite response, are expressed in terms of χ2, non-central χ2- and normal variables. The moments of Anderson's statistic are derived and used for obtaining an asymptotic normal approximation to the chance of misclassification.

UR - http://www.scopus.com/inward/record.url?scp=0347506622&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0347506622&partnerID=8YFLogxK

U2 - 10.1080/03610929808832114

DO - 10.1080/03610929808832114

M3 - Article

AN - SCOPUS:0347506622

VL - 27

SP - 577

EP - 587

JO - Communications in Statistics - Theory and Methods

JF - Communications in Statistics - Theory and Methods

SN - 0361-0926

IS - 3

ER -