Activity Number:
|
182
|
Type:
|
Contributed
|
Date/Time:
|
Tuesday, August 13, 2002 : 8:30 AM to 10:20 AM
|
Sponsor:
|
Biometrics Section*
|
Abstract - #301011 |
Title:
|
Estimating predictive MSE by cross-validation
|
Author(s):
|
Colin Mallows*+ and Lorraine Denby+ and James Landwehr
|
Affiliation(s):
|
Avaya Labs and Avaya Labs and Avaya Labs
|
Address:
|
233 Mount Airy Rd., Basking Ridge, NJ, New Jersey, 07920, USA 233 Mount Airy Rd, Basking Ridge, New Jersey, 07920, USA
|
Keywords:
|
cross validation ; calibration ; model selection ; predictive mean square error ; regression
|
Abstract:
|
Given data from a calibration experiment, we can estimate the predictive mean square error (PMSE) in two ways. First, we can fit a model to the data and use this model to derive a prediction formula and an estimate of the PMSE. Alternatively, we can use cross-validation to estimate the PMSE of this prediction formula directly. The fact that these two estimates agree gives no assurance that the model is adequate; in fact, we can have agreement even when the model is completely inappropriate.
|
- The address information is for the authors that have a + after their name.
- Authors who are presenting talks have a * after their name.
Back to the full JSM 2002 program |