Multivariate normality test based on kurtosis with two-step monotone missing data

JOURNAL OF MULTIVARIATE ANALYSIS

卷:188

Article

Kurita, Eri¹  Seo, Takashi¹ 

[1] Tokyo Univ Sci, Dept Appl Math, Tokyo, Japan

关键词: Asymptotic expansion;    Moment;    Monte Carlo simulation;    Multivariate kurtosis;    Normal approximation;   

DOI  :  10.1016/j.jmva.2021.104824

来源: Elsevier

【 摘 要 】

This paper deals with a sample measure of multivariate kurtosis, which is used as a test statistic in multivariate normality testing problems. We define a new multivariate sample kurtosis measure to provide a multivariate normality test for data with a twostep monotone missing structure. Furthermore, we derive its expectation and variance using a perturbation method. To evaluate the accuracy of a normal approximation, we conducted a Monte Carlo simulation for certain parameters. Finally, we present a numerical example to illustrate the proposed procedure. (c) 2021 Elsevier Inc. All rights reserved.

【 授权许可】

Free