|
Activity Number:
|
448
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 6, 2008 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Bayesian Statistical Science
|
| Abstract - #301130 |
|
Title:
|
A Bayesian Conceptual Predictive Statistic
|
|
Author(s):
|
Andrew Neath*+ and Joseph E. Cavanaugh
|
|
Companies:
|
Southern Illinois University Edwardsville and The University of Iowa
|
|
Address:
|
Department of Mathematics and Statistics, Edwardsville, IL, 62026-1653,
|
|
Keywords:
|
model selection ; linear regression
|
|
Abstract:
|
The conceptual predictive statistic, Cp, is widely used as a criterion for model selection in linear regression. Cp serves as an estimator of a parameter that reflects the mean squared estimation error of an approximating model. The best model within a candidate class is the one with minimum mean squared estimation error. Model selection based on Cp selects that model for which Cp is minimum as best. Strict use of the Cp criterion ignores uncertainty inherent to the model selection problem. In this paper, we present a method for handling this uncertainty using Bayesian methods.
|
