正态混合样本总体试验设计研究

Study on General Experiment Design of Normal Mixed Samples

针对正态分布类试验,从参数的估计量概率分布比对出发,采用直接推导或大规模仿真计算概率分布曲线的方法,形成正态混合样本总体试验质量评价方法。计算表明,以均值估计为目的的正态分布类试验,其样本设计与试验样本总体是否满足同一分布无关;以方差估计为目的的正态分布类试验,其样本设计与样本总体是否满足同一分布有关,并存在样本损失。

For the type of test of Normal distribution,based on the probability distribution comparison of the parameter estimators,the method of calculating the probability distribution curve by direct derivation or large-scale simulation is adopted to form the method of evaluating the quality of the normal mixed samples total test.The results show that the sample design is independent of whether the test sample population satisfies the same distribution for the normal class distribution experiments with the purpose of estimating the mean value,and for the normal class distribution experiments with the purpose of estimating the variance,the sample design is related to whether the sample population satisfies the same distribution,and there is sample loss.