Activity Number:
|
478
|
Type:
|
Contributed
|
Date/Time:
|
Wednesday, August 9, 2006 : 2:00 PM to 3:50 PM
|
Sponsor:
|
Section on Bayesian Statistical Science
|
Abstract - #306772 |
Title:
|
Bayesian Procrustes Analysis
|
Author(s):
|
Athanasios Micheas and Yuqiang Peng*+
|
Companies:
|
University of Missouri-Columbia and University of Missouri-Columbia
|
Address:
|
Department of Statistics, Columbia, MO, 65211,
|
Keywords:
|
Bayesian computation ; complex normal distribution ; Procrustes analysis ; similarity transformations
|
Abstract:
|
We introduce a Bayesian framework upon which Procrustes analysis maybe conducted. We model shapes of objects in two dimensions and develop Bayesian methods for estimation of the parameters in the similarity transformations and compare with the classic Procrustes fit. We also discuss the Bayesian methodology for a full Procrustes analysis, where a population of shapes is considered and we desire estimation of mean shape. The Bayesian approach allows us to compute point estimates and credible sets for the Procrustes fit and the Full Procrustes superimposition. The method is illustrated through an example from hydrology, where shapes of storm systems are created from radar images. We then employ precipitation forecasts and the actual truth at specific times, in order to perform verification of the forecasting method using Bayesian Procrustes methods.
|