缺失数据情形条件分位数的经验似然置信区间

在响应变量满足随机缺失机制下,采用经验似然方法和C-C方法,分别构造不含附加信息和含附加信息时条件分位数的经验似然比统计量的渐近分布,以及条件分位数的经验似然置信区间,并证明一类检验问题的渐近功效随着信息量的增加而非下降.

Empirical Likelihood Confidence Intervals for Conditional Quantiles with Missing Data

When the response variable is missing at random, empirical likelihood and the C-C method is combined to obtain asymptotic distributions of empirical likelihood ratio statistics in the absence and presence of some auxiliary information. Then empirical likelihood ratio confidence intervals for conditional quantiles are constructed. It is shown that asymptotic ef- ficacy of some test problems increases along with raise of the information quantity.