多目标调查的MPPS平衡抽样方法研究

doi:10.12341/jssms21298

系统科学与数学 ›› 2022, Vol. 42 ›› Issue (3): 715-729.DOI: 10.12341/jssms21298

多目标调查的MPPS平衡抽样方法研究

杨贵军, 沈文静, 梁鑫宇

天津财经大学统计学院, 天津 300222

收稿日期:2021-06-06 修回日期:2021-09-08 出版日期:2022-04-20 发布日期:2022-04-20
基金资助:
国家社会科学基金重点项目（20ATJ008），天津市2019年度哲学社会科学规划重点课题（TJTJ19-001），国家自然科学基金项目面上项目（11471239），内蒙古纪检监察大数据实验室2020-2021年度开放课题（IMDBD202002，IMDBD202004）资助课题.

PDF ( 41 ) 引用

杨贵军, 沈文静, 梁鑫宇. 多目标调查的MPPS平衡抽样方法研究[J]. 系统科学与数学, 2022, 42(3): 715-729.

YANG Guijun, SHEN Wenjing, LIANG Xinyu. On MPPS Balanced Sampling for Multipurpose Survey[J]. Journal of Systems Science and Mathematical Sciences, 2022, 42(3): 715-729.

On MPPS Balanced Sampling for Multipurpose Survey

YANG Guijun, SHEN Wenjing, LIANG Xinyu

School of Statistics, Tianjin University of Finance and Economics, Tianjin 300222

Received:2021-06-06 Revised:2021-09-08 Online:2022-04-20 Published:2022-04-20

平衡抽样是广受关注的利用辅助信息改善样本结构的抽样方法.MPPS抽样在多目标调查中采用的与多变量规模成比例的入样概率，可以在平衡抽样中精确满足.基于MPPS抽样和平衡抽样的性质，文章提出多目标调查下的MPPS平衡抽样，该方法主要思想是将多个调查变量的辅助信息在确定入样概率和随机抽样过程中同时使用，提高HT估计量精度.模拟分析得出MPPS平衡抽样优于等概率平衡抽样、MPPS系统抽样和MPPS泊松抽样.同时，讨论了MPPS平衡抽样HT估计量的理论性质，对于随机排列的总体，给出适用的近似方差和方差估计.最后，以我国主要畜禽调查为例，设计一套符合中国现实的多目标调查方案，用MPPS平衡抽样抽取样本，构造主要畜禽存栏量的HT估计及方差估计，为中国政府统计调查提供新选择.

关键词： MPPS抽样, 平衡抽样, 多目标调查, 辅助信息

Balanced sampling is a widely applied sampling method that uses auxiliary information to improve the sample structure. The inclusion probabilities of MPPS (Multivariate Probability Proportional to Size) sampling used in multipurpose surveys can be exactly satisfied in balanced sampling. Based on the properties of MPPS sampling and balanced sampling, this paper proposes MPPS balanced sampling for multipurpose surveys. The main idea of this method is to use the auxiliary information of multiple survey variables in determining the inclusion probability and random sampling at the same time to improve the accuracy of the estimator. Simulation analysis shows that MPPS balanced sampling is better than equal probability balanced sampling, MPPS systematic sampling and MPPS Poisson sampling. At the same time, the theoretical properties of the HT estimator of MPPS balanced sampling is discussed. For a randomly arranged population, the applicable approximate variance and variance estimation are given. At last, Chinese main livestock and poultry survey is taken as an example, and a multipurpose sampling survey scheme of MPPS balanced sampling is proposed. This could provide a new option for Chinese government statistical survey.

Key words: MPPS sampling, balanced sampling, multipurpose survey, auxiliary information

MR(2010)主题分类:

62D05

分享此文:

[1] 曾五一. 统计调查体系与调查方法问题研究. 北京:中国统计出版社, 2009. (Zeng W Y. Research on the Problem of Statistical Investigation System and Investigation Technique. Beijing:China Statistics Press, 2009.)
[2] Bailey J T, Kott P S. An application of multiple list frame sampling for multipurpose surveys. ASA Proceedings of the Section on Survey Research Methods, Anaheim, 1997.
[3] Steiner M, Bailey J T, Beranek J, et al. Improvement of sample design by using results from the agricultural census. International Conference on China Agricultural Census Results, Beijing, 2000.
[4] Kott P S, Bailey J T. The theory and practice of maximal brewer selection with Poisson PRN sampling. Second International Conference on Establishment Surveys, Buffalo, 2000.
[5] 金勇进, 栾文英. 多变量与规模成比例概率抽样的有关问题. 统计与信息论坛, 2004, (1):5-8. (Jin Y J, Luan W Y. Probability sampling of multivariate probability proportional to size. Journal of Statistics and Information, 2004, (1):5-8.)
[6] 张勇, 周巍, 涂玉娟. MPPS抽样设计方差估计的比较研究. 统计研究, 2006, (4):64-68. (Zhang Y, Zhou W, Tu Y J. Comparative study on variance estimation of MPPS sampling design. Statistical Research, 2006, (4):64-68.)
[7] 刘建平, 陈光慧. MPPS抽样下Hansen-Hurwitz估计量的扩展. 统计研究, 2005, (5):50-53. (Liu J P, Chen G H. Extension of Hansen-Hurwitz estimator under MPPS sampling. Statistical Research, 2005, (5):50-53.)
[8] Tille Y. Ten years of balanced sampling with the cube method:An appraisal. Survey Methodology, 2011, 37(10):215-226.
[9] 巩红禹,金勇进.住户调查中代表性样本的一种探索获取方法-平衡抽样设计.统计研究, 2015, 32(9):84-90. (Gong H Y, Jin Y J. Exploring representative sample in household survey-Balanced sampling. Statistical Research, 2015, 32(9):84-90.)
[10] Brewer K R W. Design-based or prediction-based inference? Stratified random vs stratified balanced sampling. International Statistical Review, 1999, 67(1):35-47.
[11] Claude J, Deville J C, Tillé Y. Efficient balanced sampling:The cube method. Biometrika, 2004, (10):893-912.
[12] Chauvet G, Tillé Y. A fast algorithm of balanced sampling. Computational Statistics, 2006, (21):53-62.
[13] 巩红禹,张士琦,王春枝. 基于分层平衡抽样的多目标代表性样本设计-以住户调查为例. 统计与信息论坛, 2021, 36(2):36-44. (Gong H Y, Zhang S Q, Wang C Z. Multi-objective representative sample exploration design based on stratified balanced sampling-A case study of household survey. Journal of Statistics and Information, 2021, 36(2):36-44.)
[14] Christine M. Use of balanced sampling in the framework of the master sample for french household surveys. Joint Statistical Meeting of the American Statistical Association, Seattle, 2006.
[15] Chipperfield J. An evaluation of cube sampling for ABS household surveys. Research Paper, Methodology Advisory Committee, Australian Bureau of Statistics, 2007.
[16] Deville J C, Till'e Y. Variance approximation under balanced sampling. Journal of Statistical Planning and Inference, 2005, 128(2):569-591.

[1]	胡桂华, 叶宝红, 漆莉. 基于双系统估计量的人口普查净误差估计[J]. 系统科学与数学, 2022, 42(3): 676-696.
[2]	崔铁军，马云东. 离散型空间故障树构建及其性质研究[J]. 系统科学与数学, 2016, 36(10): 1753-1761.
[3]	赵亮，杨战平. 一种针对实验数据的裕量和不确定性量化方法[J]. 系统科学与数学, 2016, 36(8): 1138-1149.
[4]	崔铁军，马云东. 基于多维空间事故树的维持系统可靠性方法研究[J]. 系统科学与数学, 2014, 34(6): 682-692.
[5]	洪志敏，闫在在. 估计敏感特征比例和敏感指标均值的问卷调查技术[J]. 系统科学与数学, 2012, 32(3): 344-354.
[6]	于丹;李学京;姜宁宁;杨军. 复杂系统可靠性分析中的若干统计问题与进展[J]. 系统科学与数学, 2007, 27(1): 68-81.
[7]	邹国华;冯士雍. 超总体模型下有限总体的估计[J]. 系统科学与数学, 2007, 27(1): 27-38.
[8]	刘礼;邹国华. 缺失数据下Jackknife方差估计量的渐近设计无偏性[J]. 系统科学与数学, 2006, 26(4): 491-503.

阅读次数

全文

摘要