论文部分内容阅读
本文提出两个新的估计量，利用观察数据中的总体辅助信息来估计有限总体分布函数，并通过两个人工总体的模拟实验，比较新的估计量、传统的估计量及Rao，Kover＆Mantel（,1990）提出的估计量的相对平均误差与相对标准差。结果表明，从相对标准差的角度分析，两个新的估计量有一个是四个估计量中精度最好的一个，另一个也有很好的表现；而且它们在模型有所偏差时都具备了较好的稳健性。“,”In this paper, we offer two new estimators for the finite population distribution function when there is auxiliary population information available from survey data. The relative mean errors and relative root mean square errors of these estimators are compared with a customary design-based estimator and an estimator of Rao, Kover ＆ Mantel（1990） through a simulation study on two artificial populations. Results are reported that one of new estimators is the thorough winner in the four design- based estimators in terms of relative root mean square error and the other performs well. Moreover, the two new estimators are quite robust under model misspecifications.