zbMATH — the first resource for mathematics

An almost unbiased ridge estimator. (English) Zbl 0687.62054
Summary: An estimator for the regression parameters in a multiple linear regression model is proposed by jack-knifing the ridge estimator. When the matrix (X’X) is of diagonal form, the resulting estimator reduces the bias uniformly in all components of the parameter vector. However, in the general case we can only guarantee that the sum of squares bias is reduced. The jackknife procedure naturally lends itself in rendering simple confidence intervals for the parameter vector, which is outlined.

62J07 Ridge regression; shrinkage estimators (Lasso)
62J05 Linear regression; mixed models