基于线性回归变量误差模型的工具变量法与校正似然法的比较  

Comparison of Instrumental Variable Estimation and Corrected Likelihood Method Based on Linear regression measurement error models

作　　者：关静[1] 陈永沛 

机构地区：[1]天津大学数学学院,天津300350

出　　处：《统计与决策》2017年第10期81-84,共4页Statistics & Decision

摘　　要：文章介绍了线性回归变量误差模型参数估计的两种方法——工具变量法和校正似然法,然后通过数值模拟的方式对这两种方法的估计结果进行比较,说明这两种方法在不同假定下估计的优劣,最后通过实例计算来进行验证,并得到一些有用的结论。This paper introduces two methods of parameter estimation of linear regression measurement error models--Instrumental Variable Estimation and Corrected Likelihood Method. And then numerical simulation is given to compare the estimation results of the two methods. Advantages and disadvantages of these two methods under different assumptions are also described in the paper. Finally some useful conclusions are obtained through the case calculation and verification.

关 键 词：线性回归 变量误差模型 工具变量法 校正似然法 

分 类 号：O212[理学—概率论与数理统计]