非齐次等式约束线性回归模型的条件岭估计的效率

The Relative Efficiency of the Conditional Ridge Estimation of Parameter in Restricted Linear Regression Model

利用文[7]提出的相对效率定义,研究文[8]、[9]非齐次等式约束线性回归模型回归系数的狭义条件岭估计和广义条件岭估计的效率,证明在一定条件下,狭义条件岭估计和广义条件岭估计的效率比约束最小二乘估计的效率高,并且广义条件岭估计的效率比狭义条件岭估计的效率高,从而进一步证明了在病态的模型中,有偏估计优于无偏估计。

Research the relative efficiency of the conditional ridge estimation of regression Coefficient in restricted lin ear regression model by the definition of the relative efficiency , Prove the relative efficiency of the conditional ridge estimation are not lower than that of the least squares estimation,and the relative efficiency of the generalized is not lower than that of ridge estimation, the biased estimation is better than the unbiased estimation in the complex collinear model.