本文考虑带有不完全信息随机截尾试验模型,当模型中寿命随机变量的总体分布类型完全未知时,针对总体均值型参数进行统计推断,构造经验对数似然比统计量,模拟表明该统计量的分布与标准卡方分布拟舍得很好,由于数据是不完全数据,借鉴[11]采用广义EM算法计算该统计量,数值模拟表明这是个行之有效的方法.
We consider the distribution of variable of interest in random censoring model with incomplete information is unknown completely. We first apply the empirical likelihood method to the model. In this situation,we also use generalized EM algorithm to compute the proposed empirical likelihood ratio where the constraint is of mean type. Finite sample performance of the proposed methods is illustrated in a simulation study. Tests based on empiri- cal likelihood ratio performs well.