摘要
将多元线性模型 Y=XB+E,[E～(0,σ~2V(?)I)]拉直得出一元线性模型,从而得出多元线性假设检验的UMPI解,不便之处是V不知道,文中得出良好的渐近结果,有效地用于泛假设检验问题,可据以得统计聚类分析解。文中论证是在(?)球等高分布基础上进行的。
The multivariate linear models Y=X B+E, where E～(0, V(?)I), may be straightened to be the ordinary linear models. Thus a certain kind of UMPIT of the multivariate linear hypothesis may be obtained. The inconvenience is that V would be unknown. But we get a good asymtotic result in such case, and it has been aPPlied effectively in the problem of universal hypothesis. Accordingly, a statistical model of cluster analysismay be derived.
出处
《暨南大学学报（自然科学与医学版）》
CAS
CSCD
1989年第1期1-8,共8页
Journal of Jinan University(Natural Science & Medicine Edition)
关键词
多元线性模型
泛假设
椭球等高分布
Multivariate linear model, Universal hypothesis, Elliptically contoured distribution
