|
Activity Number:
|
139
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, August 4, 2008 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Statistical Computing
|
| Abstract - #302597 |
|
Title:
|
Far Casting Cross Validation
|
|
Author(s):
|
Patrick Carmack*+ and Jeffrey Spence and Qihua Lin and William R. Schucany and Richard Gunst
|
|
Companies:
|
The University of Texas Southwestern Medical Center and The University of Texas Southwestern Medical Center and The University of Texas Southwestern Medical Center and Southern Methodist University and Southern Methodist University
|
|
Address:
|
5323 Harry Hines Blvd., Dallas, , TX, 75390-8896,
|
|
Keywords:
|
model selection ; correlated data ; extrapolation
|
|
Abstract:
|
Cross validation has long been used for choosing tuning parameters and other model selection tasks. It generally performs well provided the data are independent, or nearly so. Improvements have been suggested which address ordinary cross validations (OCV) shortcomings in correlated data. While these techniques have merit, they can still lead to poor model selection in correlated data or are not readily generalizable to high dimensional data. The proposed solution, far casting cross validation (FCCV), addresses these problems. FCCV withholds correlated neighbors during cross validation, but uses the full data set once a model is selected. The result is a technique that stresses a fitted model's ability to extrapolate rather than interpolate. This generally leads to better model selection in correlated data sets.
|
