论文部分内容阅读
流程性材料最大的特点是其变异性小。对于这类总体,现有的标准差估计方法由于既包含组间差异,又包含组内差异,常常会夸大其估计误差。针对此,首先通过抽样设计,得到具有分层抽样特点的样本;然后借鉴单值—移动极差控制图中标准差的估计方法,构造了这类总体的标准差估计量。这样构造的标准差估计量,由于其估计误差中仅包含组内方差平均水平,从而更符合该类总体变异性小的特点。实际应用表明,该标准差估计量能显著降低估计误差。
The biggest characteristic of process materials is their low variability. For such aggregates, the existing standard deviation estimation method often exaggerates its estimation error because it contains both inter-group differences and intra-group differences. In view of this, we first obtain the samples with the characteristics of stratified sampling through the sampling design. Then, we estimate the standard deviation of the single value-moving-range control chart by using the standard deviation method. The standard deviation estimator constructed in this way is more suitable for this kind of population with less variability due to the fact that its estimation error only contains the average variance within the group. The practical application shows that the standard deviation estimator can significantly reduce the estimation error.