Ridge Regression for Nonstandardized Models.
Abstract
In the usual regression model y = Xbeta + epsilon with X of full rank and epsilon approx. N(O, sigma squared I), the ordinary least squares estimator beta = (X'X) inverse X'y has extremely large mean square error when X is an ill-conditioned matrix. This paper compares ridge estimators for beta that arise when the biasing factor (k) is applied at different stages of standardization (i.e., centering and scaling), and shows which estimators are identical and which are different. In addition, results of a small-scale simulation are discussed.
Document Details
- Document Type
- Technical Report
- Publication Date
- Oct 01, 1977
- Accession Number
- ADA045406
Entities
People
- Dennis E. Smith
- Terry L. King