针对不完全样本观测数据,讨论了一类均匀分布总体参数的区间估计问题.利用样本中位数给出了构造置信区间的一个新枢轴量,推导出了枢轴量的概率密度函数表达式,并且在大样本场合,讨论了总体参数的近似置信区间.该方法不仅适用于不完全数据场合,而且还适用于样本中可能存在异常数据的情形,具有稳健性.
Based on data observed from incomplete samples, the problem of confidence interval estimation is investigated for uniformly distributed population parameter. By use of median of the samples, a new pivot variable is given for the construc- tion of the confidence interval and deduced from which is the probability density function of the pivot quantity. An approximate confidence interval from large sample is put under discussion as well. The method is applicable not only to the cases of incom- plete data but also to cases when there exist outliers in the samples.