带AR(1)误差线性回归模型中杠杆点的度量与影响分析

Measurement and Impact Analysis of Leverage Points in Linear Regression Model with AR(1) Error

杠杆点的度量与影响分析是回归诊断中的一个重要问题.本文通过构建投影矩阵和拟投影矩阵,对带AR(1)误差且存在复共线性的线性回归模型中广义最小二乘估计、主成分估计、r-k类估计、r-d类估计杠杆点的度量和影响进行了模拟分析与实证分析,并对各估计杠杆点的度量值做了比较研究.

The measurement and impact analysis of the Leverage points is an important issue in the regression diagnosis.In this paper,by constructing projection matrix and quasi-projection matrix,the measurement and influence of generalized least squares estimator,principal component estimator,r-k class estimator,r-d class estimator of leverage points in the linear regression model with AR(1)error and multicollinearity are simulated and empirically analyzed,and the measurement values of each estimators leverage are compared.