复共线性条件下广义岭估计的改进

Research on improvement of generalized ridge estimation in multicollinearity condition

对于复共线性条件下线性回归模型的广义岭估计进行了进一步的研究。针对线性回归模型病态的根本原因,提出了一类新的估计——0-K型广义岭估计。研究这一估计的性质,证明利用0-K型广义岭估计技术可以改进广义岭估计(在均方残差意义下)。文中的方法为病态线性回归模型系数的有偏估计提供了改进的技术途径。

Further research on the generalized ridge estimation of linear regression model in multicollinearity condition is done.In the light of the essence of the ill condition in the linear regression model,this paper first proposes the 0-K class of estimators of the coefficients.The 0-K class of estimators if studied,it will be proved that under the mean square residua criterion the estimators can be improved via the generalized ridge regression technique.The method proposed in this paper provides a technical way ...