|
Activity Number:
|
29
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Sunday, July 29, 2007 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Physical and Engineering Sciences
|
| Abstract - #308274 |
|
Title:
|
The Efficiency of Ridge Regression Estimators with Respect to Zellner's Balanced Loss Function
|
|
Author(s):
|
Marvin Gruber*+
|
|
Companies:
|
Rochester Institute of Technology
|
|
Address:
|
85 Lomb Memorial Drive, Rochester, NY, 14623,
|
|
Keywords:
|
balanced loss function ; matrix loss ; simultaneous estimation ; total mean square error ; ridge estimator
|
|
Abstract:
|
The optimum ridge estimator is obtained for a setup with r linear models for a generalization of Zellner's balanced loss function. Two important special cases are considered, matrix loss and the total mean square error. For each of these two cases an estimator of the biasing parameter matrix is obtained that produces a mean square error that is uniformly smaller than that of the least square estimator. The results are illustrated with an example that uses numerical data. In general the best improvement in the MSE over the least square estimator occurs for the classical loss function. Parameters with smaller values produce smaller MSE.
|
