辅以排序集样本的分层抽样方法研究  

The Research of Stratified Sampling Based on Ranked Set Sample

作　　者：张建军[1] 乔松珊[2] 

机构地区：[1]河南农业大学信息与管理科学学院,河南郑州450002 [2]中原工学院信息商务学院,河南郑州450007

出　　处：《统计与信息论坛》2012年第5期14-18,共5页Journal of Statistics and Information

基　　金：河南省教育厅人文社科研究项目<利用辅助信息抽样研究工业用电定额策略>(2011-QN-220)

摘　　要：分层抽样中,样本在各层中的不同获取方式会对估计量的精度和试验费用产生一定的影响,而已有的理论方法大多不能在提高精度的同时降低调查费用。为此,将排序抽样与分层抽样方法相结合,提出了辅以排序集样本的分层抽样方案,并得到了总体均值的估计量以及这一估计量的良好性质。这些结果表明,与单一的分层随机抽样相比,这种抽样设计的估计量具有更高的精度,同时也节约了各层抽样调查的费用。In stratified random sampling,the different ways of sampling in different strata can influence the precision of the estimator and the cost of the investigation.And the existing theory methods are mostly not in both improving the accuracy and reduce the cost of the investigation,In this paper,we combined the ranked set sampling and the stratified random sampling and considered a new sample method.We estimated the population mean using the stratified rank samples,it is proved that this estimation is more efficient compared with the stratified random samples in the estimate relative precision,At the same time,the method saved the cost of the investigation.

关 键 词：分层抽样 排序集抽样 总体均值 估计相对精度 调查费用 

分 类 号：O212.2[理学—概率论与数理统计]