双样本的协方差齐性检验及其在刑事技术中的应用

Equality test of two covariance matrices for twosample Problem and it′s application in forensic science

文献［３］给出了双样本协方差矩阵检验的统计量λ及其概率密度函数ｐ（λ），但由于ｐ（λ）太复杂，其分位数表尚未算出，无法实用；本文采用了统计模拟的方法给出λ的经验分布，由此计算出λ的上１００α％分位数表，最后给出了刑事技术中的两个应用实例。

AbstractThe paper by C.G.Khatri and M.S.Srivastara, On Exact Non-null Distribution of Likelihood Ratio Criteria for Sphericity Test and Equality of Two Covariance Matrices (Sankhya A 33, 201-206) described statistic in test of covariance matrices of two samples, and provided its density function p(λ). Since the density function is too complex, whose significance point table has not been given, it cannot be used in practice. The present paper makes use of simulation and calculates the experience distribution of λ. On the base of this, gives the tables of under 100% significance point. Two practice examples in forensic technique are made.