Is Cross-Validation the Best Approach for Principal Component and Ridge Regression?

摘要

A recent study by Frank and Friedman (1993) indicated thatrncross-validated ridge regression performed well whenrncompared to partial least-squares regression and crossvalidatedrnprincipal components regression. Thorpe andrnScharf (1995) consider a number of uncross-validated ridgetypernestimators from an engineering point of view. In thisrnpaper we examine a variety of estimators to see if we can dornas well as or nearly as well as fully cross-validated ridgernregression. We conclude that when the number ofrnparameters does not exceed the number of observations, itrnmay be possible to avoid cross-validation.