摘要
本文研究了多元分布的正态性检验问题,用投影寻踪自助法,获得了投影偏度和投影峰度正态性检验统计量,证明了在零假设成立时,所提出的偏度和峰度检验统计量的极限分布为一高斯过程的上界.为计算机模拟计算提供了有力的手段和依据.
An approach for testing multinormality based on the PP-skewness and PP-kurtosis test by using PP method and Bootstrap method is proposed. The asymptotic distributions of those test statistics under null hypotheseis are obtained, It is proved that they have the same asymptotic distribution which is the supremum of a Gauss process.
出处
《数学杂志》
CSCD
北大核心
2006年第2期147-154,共8页
Journal of Mathematics
基金
国家自然科学基金资助项目(19771011).
