线性测量误差模型中复合分位数估计的随机加权方法  

Random weighted method to estimate composite quantile for linear errors-in-variables models

作　　者：葛悠美 刘莹[1] 姜荣[1] 

机构地区：[1]东华大学理学院,上海201620

出　　处：《黑龙江大学自然科学学报》2015年第2期154-162,共9页Journal of Natural Science of Heilongjiang University

基　　金：国家自然科学基金资助项目(11171062);上海市教委科技创新重点项目(12ZZ063)

摘　　要：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

关 键 词：线性测量误差模型 渐近正态性 随机加权 复合分位数回归 

分 类 号：O436[机械工程—光学工程]