多元正态性检验三种方法的比较及SAS程序设计
Comparison of three methods for multivariate normality test and SAS program design
朱宁 1赵肖肖 1唐庆华1
作者信息
- 1. 桂林电子科技大学数学与计算科学学院,广西桂林541004
- 折叠
摘要
讨论了正态性检验的三种主要方法:二元等概椭圆检验法、多元x2统计量的Q-Q图检验法和主成分检验法,分别介绍了三种方法的基本原理及步骤；并利用统计软件SAS设计出三种检验方法的程序；最后比较分析三种方法的应用场合,得到主成分检验法具有较好的检验效率.
Abstract
The article discussed three kinds of main methods of multivariate normality test; two di-mension equal probability elliptic test method, multivariate chi-square Q-Q plot test method and prin-cipal component method, respectively introduced the basic principle and steps of these methods, and using statistical software SAS designed three test methods of program, finally analyzed their applica-tions. Through the research on the inspection, we obtained that the principal component method has a good test efficiency.
关键词
多元正态/等概椭圆法/卡方统计量/Q-Q图/主成分/马氏距离/SASKey words
multivariate normality/equal probability elliptic method/chi-square statistics/Q-Q plot/principal component/Mahalanobis distance/SAS引用本文复制引用
出版年
2012
NETL
NSTL
维普
万方数据