讨论了两个总体协方差差异的检验问题,给出了该检验问题的U统计量.该检验方法不仅适合参数统计结构,也适用于非参数统计结构,与经典的Pearson矩相关系数检验法相比,其渐近相对效率为1.
The testing problem about the difference of the covariance in two populations is discussed. A U-statistics is given about this testing problem. This method is suitable for parametric and non-parametric statistical structure. Compared with the classic Pearson moment correlation coefficient testing method, its asymptotic relative efficiency is 1.