二维平衡单水平轮换设计及其广义组合估计方法研究  

The Study of Two-way Balanced One-level Rotation Design and Generalized Composite Estimation Methods

作　　者：陈光慧[1] 刘建平[1] 

机构地区：[1]暨南大学经济学院,广东广州510632

出　　处：《数理统计与管理》2011年第4期696-704,共9页Journal of Applied Statistics and Management

基　　金：国家社会科学基金青年项目(10CTJ006);全国统计科学研究重大项目(2009LD001)资助

摘　　要：针对国内外现行单水平轮换调查存在的一系列问题,本文提出二维平衡单水平轮换模式设计及其广义组合估计方法。在这套设计方法中,将轮换偏差和两类相关关系引入到广义组合估计量公式中,提高了估计量的准确性,同时在连续性的多目标调查下还应用最优化方法确定一套最优权重系数,使得抽样调查的整体精度达到最高,同时又能够调整权重,满足各类目标量精度高低的要求。这套系统化、理论化的设计方法将为我国今后更好地开展各类单水平轮换调查提供理论指导。This paper puts forward two-way balanced one-level rotation scheme design and generalized composite estimation method according to a series of problems from the existing one-level rotation surveys. Under this set of design and method, rotation bias and two types of correlation are introduced into the generalized composite estimator which improves the accuracy of estimation. At the same time, optimization method is applied to determine one set of the best coefficients under the successive multi- objective surveys which not only makes the whole accuracy of sampling estimation reach the highest, but also takes into account the accuracy requirements of various types of target estimators. This set of systematic and theory-based design and method would provide theoretical guidance for carrying out the one-level rotation surveys in the future China

关 键 词：连续性抽样调查 平衡单水平轮换 广义组合估计量 轮换偏差 

分 类 号：C811[社会学—统计学] O212[理学—概率论与数理统计]