An Alternative Approach for Selecting Ridge Parameter for Ordinary Ridge Regression Estimator

Hazim Mansoor Gorgees, Fatimah Assim Mahdi

Abstract: In the presence of multicollinearity, the parameter estimation method based on the ordinary least squares procedure is unsatisfactory. In 1970, Hoerl and Kennard introduced alternative method distinguished as ridge regression estimator. In such estimator, ridge parameter or biasing constant plays an important role in estimation. Various methods were suggested by many researchers for choosing the ridge parameter. In this article we employed the concept of condition number to suggest a new method for selecting the ridge parameter. The performance of the proposed method is assessed and compared with other traditional methods through simulation study in terms of mean square error (MSE). The method developed in this paper seems to be reasonable since it has smaller MSE than the other stated methods

Keywords: multicollinearty, ridge regression, ridge parameter, singular value decomposition, condition number