二维正态总体相关性的假设检验

摘要

在样本相关系数具有渐近正态性的条件下,给出了一种二维正态总体分量相关性的假设检验方法,并且与大样本情形下的传统检验方法比较了功效,得到渐近相对效率为1的结论,且新的检验统计量形式明显简单,从而说明新的检验方法是简便可行的.%A hypothesis testing of the dependency of two-dimensional normal population is given,under the condition of the sample correlation coefficient with asymptotic normality.Compared with the traditional testing method the asymptotic relative efficiency is 1.The new test statistic is simple,and the new testing method is proved to be easy and feasible.