给出线性测量误差模型中随机加权复合分位数估计,用随机加权方法逼近复合分位数回归的分布,并证明这种逼近是以概率1渐近有效地。模拟研究以及艾滋病的临床试验数据验证了方法的有效性。
The random weighted composite quantile estimation in linear errors-in-variables model is giv- en. The random weighted method is used to approximate the distribution of the composite quantile re- gression, and it is demonstrated that the approximation is asymptotically valid with probability one. Fi- nally, simulated examples and AIDS real data are given to illustrate the effectiveness of the proposed method