Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
1151333 | Statistics & Probability Letters | 2016 | 5 Pages |
Abstract
Two data-driven procedures, based respectively on the L-curve and generalized cross-validation, are proposed for ridge regression under Aalen’s additive risk model. Monte Carlo simulations show that the L-curve is a useful criterion for identifying a nominal degree of regularization that appreciably reduces variance, particularly in smaller samples.
Related Topics
Physical Sciences and Engineering
Mathematics
Statistics and Probability
Authors
Audrey Boruvka, Glen Takahara, Dongsheng Tu,