设有两个非参数总体,其样本数据不完全,用分数填补法补足缺失数据,得到两总体的“完全”样本数据,在此基础上构造两总体分位数差异的经验似然置信区间.模拟结果显示,分数填补法可以得到更加精确的置信区间.
Consider two independent nonparametric populations with missing data. Frac- tional' imputation method is used to impute the missing data to obtain "complete" data. Empirical likelihood confidence intervals for quantile difference of two populations are then constructed based on the fractional imputation. It is shown that fractional imputation can improve the accuracy of confidence intervals.