|
Activity Number:
|
460
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Wednesday, August 9, 2006 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Statistics and the Environment
|
| Abstract - #305623 |
|
Title:
|
Modified Power Priors with Multiple Historical Datasets in Water Quality Evaluation
|
|
Author(s):
|
Yuyan Duan*+ and Keying Ye and Eric P. Smith
|
|
Companies:
|
Bristol-Myers Squibb Company and The University of Texas at San Antonio and Virginia Polytechnic Institute and State University
|
|
Address:
|
22 Brandywine Court, Washington Crossing, PA, 18977,
|
|
Keywords:
|
historical data ; power prior ; water quality standards
|
|
Abstract:
|
The modified power prior has been proposed to utilize additional information to improve the water quality evaluation. In this presentation, we investigate three power prior methods used in incorporating multiple historical datasets. One of the methods is chosen among the three by comparing the mean squared errors (MSE) in estimating the parameter of interest. The Bayesian statistical model using the selected power prior method is compared with the random effects model for the accommodation of potential heterogeneity between current and historical samples. Furthermore, we apply the modified power prior with a random effects model. This approach is shown to recognize the differences among groups the most, compared to using only power priors and only random effects models. Several sets of water quality data are studied to illustrate the implementation of the modified power prior method.
|
