等级缺失数据下两正态协差阵的推断  

Inferences of Two Normal Covariance Matrices Under Grade Missing Data

作　　者：禹建奇[1] 王想[1] Yu Jianqi;Wang Xiang(College of Science,Guilin University of Technology,Guilin Guangxi 541004,China)

机构地区：[1]桂林理工大学理学院,广西桂林541004

出　　处：《统计与决策》2020年第19期38-41,共4页Statistics & Decision

摘　　要：文章考虑两协差阵相等的检验问题。假设数据为多元正态且为等级缺失形态,提出了一个类似于似然比检验的统计量,并求出了该统计量的近似分布,从而得到检验方法。蒙特卡洛数据模拟证明,即使在中等程度的小样本上,该近似结果也非常满意。最后用一个例子说明了分析方法的应用。This paper considers testing equality of two covariance matrices.Assuming that the data are multivariate normal and of hierarchical deletion pattern,the paper proposes a statistic quantity similar to the likelihood ratio test,and then solves the approximate distribution of the statistics,thus obtaining the test method.Monte Carlo data simulation indicates that this approximation is satisfactory even for a small sample of moderate degree.Finally,an example is given to illustrate the application of the analysis method.

关 键 词：不完全数据 协差阵 检验水平 

分 类 号：O213.2[理学—概率论与数理统计]