Truncation Ridge Estimation Based on the Biased-correctedTheory by Multicollinearity Diagnosis

In order to avoid the bad effect provided by the biases of the biased estimation,bias-corrected regu-larization is presented.But how to choose the bias-corrected term is a key link.In this paper,based on the keyinformation from the multicollinearity diagnosis,measurement and test,parameters seriously harmed by themulticollinearity are estimated.Then based on the bias-corrected theory,the analytical condition of the bias-corrected regularization is derived.Truncation ridge estimation based on the bias-corrected theory by multicol-linearity diagnosis is proposed.The numerical results demonstrate that the new method can improve the nu-merical stability and accuracy of parameter estimation.