广义线性度量误差模型

General Error-in-Variable Linear Models

在线性度量误差模型中,需要假设所有变量的观测值都含有未知度量误差．因而该模型不适用于一部分变量的观测值含有度量误差、而另一部分变量的观测值可精确得到的情况．为此,本文提出了广义函数、结构和超结构关系线性度量误差模型．进一步,这里还讨论了这些广义线性度量误差模型中参数的最小二乘和极大似然估计方法,给出了参数估计的表达式．

In linear error-in-variable models, it is supposed that the observed values of all variables include unknown measurable errors. So the models can not be used to describe the linear relationship of variables in which some variables can be observed exactly, and others can not be observed exactly. Therefore, we offer general error-in-variable models to solve the problem, and discuss the least squares method and likelihood method for the general models.