基于约束最小二乘估计的影响分析

Influence analysis based on constrained least squares estimator

研究了一类约束条件下的线性回归模型。首先,用Lagrange乘子法得到了刀切约束最小二乘估计;然后,定义了用于影响分析的Cook统计量,并得到了近似Cook统计量;最后,用算例说明了近似Cook统计量的有效性。

In the paper,we investigate a linear regression model with constraint conditions.Firstly,we obtain jackknife constrained least squares estimator by the Lagrange multiplier method.Secondly,we give the definition of Cook＇s statistic for influence analysis,and obtain the approximate Cook＇s statistic.At last,by an example,it is shown that the approximate Cook＇s statistic is valid.