摘要
对于测量误差模型,协变量和响应变量真值之间可能不存在完全匹配关系,引入方程误差可以解决此类数据建模问题.本文针对重复测量数据,研究测量误差模型的建模和估计问题,给出含有方程误差时参数估计的EM算法,同时获得基于Fisher信息阵的参数估计渐近协方差矩阵.模拟计算和实例分析的结果说明了本文估计方法的有效性和考虑方程误差的重要性.
In some measurement error （ME） models, the true variables may be not perfectly related. In this case, the equation error should be considered into the models. In this work, we propose a new replicated structural ME model with equation error. Maximum likelihood estimates are computed via an EM type algorithm method. A closed expression is presented for the asymptotic covariance matrix of those estimators based on the Fisher information matrix. The results of simulation studies and a real data set analysis confirm the effect of the estimation method and illustrate the importance of the equation error.
出处
《应用数学》
CSCD
北大核心
2013年第3期627-634,共8页
Mathematica Applicata
基金
国家自然科学基金(11171065)
江苏省自然科学基金(BK2012459)
江苏省大学生实践创新训练计划重点项目(2012JSSPITP0795)
关键词
测量误差
重复测量
方程误差
EM估计
Measurement error
Replication
Equation error
EM estimator
