A consistent test for multivariate normality based on the empirical characteristic function
LetX1,X2, …,Xn be independent identically distributed random vectors in IRd,d ⩾ 1, with sample mean\(\bar X_n \) and sample covariance matrixSn. We present a practicable and consistent test for the composite hypothesisHd: the law ofX1 is a non-degenerate normal distribution, based on a weighted integral of the squared modulus of the difference between the empirical characteristic function of the residualsSn−1/2(Xj −\(\bar X_n \)) and its pointwise limit exp (−1/2|t|2) underHd. The limiting null distribution of the test statistic is obtained, and a table with critical values for various choices ofn andd based on extensive simulations is supplied.
Unable to display preview. Download preview PDF.
- Baringhaus L, Danschke R, Henze N (1988) Recent and classical tests for normality — a comparative study. Comm Statist B Comp Simul (submitted)Google Scholar
- Fisher RA (1936) The use of multiple measurements in taxonomic problems. Ann Eugenics 7/II:179–188Google Scholar
- Mardia KV (1980) Tests of univariate and multivariate normality. In: Krishnaiah PR (ed) Handbook of statistics, I Analysis of variance. North-Holland, Amsterdam, pp 279–320Google Scholar