AN ADAPTIVE GENERAL COMBINED RESTRICTED RIDGE REGRESSION ESTIMATOR FOR LINEAR REGRESSION MODEL UNDER MULTICOLLINEARITY

Abstract

The introduction by Hoerl and Kennard (1970) of a ridge regression estimator to deal with the problem of multicollinearity in regression analysis has been followed by a number of research articles in the statistics literature. The defined a class of estimators characterized by a scalar called Ridge parameter. By reducing the linear regression model to its canonical form, they proposed the general Ridge regression estimator.