基于样本均值的最优无重叠k-序对排序集抽样

发布时间：2018-07-09 14:10

本文选题：无重叠κ-序对排序集抽样 + 相对效率　；参考：《数学学报(中文版)》2017年06期

【摘要】：本文提出了无重叠κ-序对排序集抽样方法,即在每个排序集中对κ-序对个体进行观测,并且不同的排序集的κ-序对之间没有任何重复.我们首先探究了此抽样方法得到的样本均值的有效性随每个排序集中κ-序对个体间的相关性变化的趋势.κ-序对个体间的相关性越强,样本均值的有效性损失越大.本文的目的是找到无重叠κ-序对排序集抽样方法中κ-序对分配的最优方案从而使样本均值的有效性损失最小,并证明了最优的无重叠κ-序对排序集抽样比广义排序集抽样以及简单随机抽样更有效.尽管无重叠κ-序对排序集抽样方法的统计效率低于经典的排序集抽样,但是在成本模型下,最优的无重叠κ-序对排序集抽样方法可以比经典的排序集抽样更有效.
[Abstract]:In this paper, we propose a sampling method of unoverlapped 魏 -order pairs of sorting sets, in which individuals are observed in each sequence set, and there is no duplication between the 魏 -order pairs of different sorting sets. We first investigate the trend of the validity of the sample mean obtained by this sampling method with the variation of the correlation between individuals in each sequence set 魏-order. The stronger the correlation between individuals is, the greater the loss of validity of the mean sample is. The purpose of this paper is to find the optimal scheme of 魏-order pair allocation in the sampling method of unoverlapped 魏-order pairs, so as to minimize the loss of validity of the mean value of the sample. It is proved that the optimal non-overlapping 魏 -order sampling is more effective than generalized sorting set sampling and simple random sampling. Although the statistical efficiency of the unoverlapped 魏 -order-pair sampling method is lower than that of the classical sorting set sampling method, under the cost model, the optimal non-overlapping 魏 -order-pair sorting set sampling method can be more effective than the classical sorting set sampling method.
【作者单位】：上海财经大学统计与管理学院;北京大学数学学院;
【基金】：国家自然科学基金委重大研究计划重点项目(91546202)
【分类号】：O212.2

【相似文献】