论文部分内容阅读
物证检验问题可以转化为双样本均值向量的检验和协方差矩阵的检验之统计理论问题 .对于检验统计量 T2、λ3,虽然 Hotelling对均值向量的检验给出了具有许多优良性质的 T2统计量 ,但它使用的前提条件是现场和嫌疑人的样本量必须同时大于特征数 ,这在物证检验中很难满足 ,本文的拓宽了使用条件 ,使其在物证检验中能够广泛使用 ;本文采用了统计模拟的方法构造了检验均值向量和协方差阵同时相等的统计量λ4 的上 1 0 0α%分位数表 ,这就为物证检验提供了新工具 ,将对于揭露和打击犯罪、保障国家的安全与社会的稳定具有重大的社会效益 .最后给出了应用实例 .
The problem of material evidence verification can be transformed into the statistical theory of test of two-sample mean vector and test of covariance matrix.For the test statistic T2, λ3, although Hotelling’s test of mean vector gives T2 statistic with many excellent properties, However, the prerequisite for its use is that the sample size of the scene and the suspect must be greater than the number of features at the same time, which is difficult to meet in the material evidence verification. This article broadens the conditions of use and makes it widely used in the material evidence verification. The simulation method constructs a table of the upper 10 0% quantile of λ4, which is equal to the mean of the test mean value vector and the covariance matrix, which provides a new tool for material evidence verification that will be of great help in exposing and cracking down crimes and safeguarding national security And social stability has great social benefits.Finally, the application examples are given.