带有测量误差的线性回归模型

The Linear Regression Model with Measurement Error

通常在研究线性回归的问题中只是考虑因变量存在误差或扰动项.在很多情况下,线性回归中自变量也是带有误差的,并且当自变量误差存在时,用常用的参数估计方法,比如一般最小二乘法,得到的模型参数的估计值是有偏的且非一致的.采用工具变量法来处理这类问题,并通过数据模拟分析工具变量法.

Usually in linear regression problem, it＇s considered that the dependent variable has errors or disturbance. In many cases, independent variables are also with errors in linear regression model and parameter the estimators of parameters by usual methods, such as OLS, are biased and inconsistent in this case. This kind of problem is dealed by using the instrumental variable method, and this method through the data simulation is analyzed.