【摘要】 本文研究了不放回追加策略,包括基本设计和域追加设计都为简单随机抽样、分层随机抽样情形下不放回样本追加时域的估计的问题。根据不同的抽样设计给出单元的一阶及二阶包含概率的具体计算公式,并构造总体总量和域总量的Horvitz—Thompson型估计,然后基于简单随机抽样的不放回追加抽样方案,给出总体单元的前两阶包含概率。及该方案在分层抽样下的推广,在有辅助信息可用时构造域总量的分层联合比估计,并给出其方差和方差估计公式,同时我们给出了模拟结果,从模拟结果可以看出,给出的方差估计是估计量方差的近似无偏估计。更多还原

【Abstract】 This paper develops complementing strategies of sampling without replacement. Various inclusion probabilities are universally formulated for general complementing without replcement. Actually,the relevant formulae describe a general clue. Consequently,investigated in detail are several concrete complementing schemes based on simple random sampling,stratified sampling. Inclusion probabilities are specified to calculate Horvitz—Thompson estimates of universe and domain totals,formulate relevant variances and the estimators of variances. Especially,the combined ratio estimator of domain total is investigated under stratified complemental sampling when auxiliary information can be used. Numerical studies show that the proposed estimators can gain much precision.更多还原