Activity Number:
|
250
|
Type:
|
Contributed
|
Date/Time:
|
Tuesday, August 13, 2002 : 2:00 PM to 3:50 PM
|
Sponsor:
|
Section on Statistics & the Environment*
|
Abstract - #301007 |
Title:
|
Using Bayesian Model Averaging with Canonical Correspondence Analysis
|
Author(s):
|
Eric Smith*+ and Ilya Lipkovich
|
Affiliation(s):
|
Virginia Polytechnic Institute and State University and Virginia Polytechnic Institute and State University
|
Address:
|
401C Hutcheson Hall, Blacksburg, Virginia, 24061-0439, USA
|
Keywords:
|
multivariate analysis ; environmental stress ; species data ; model evaluation ; graphical assessment ; correspondence analysis
|
Abstract:
|
Canonical correspondence analysis is a variant of canonical correlation analysis that is used in environmental studies to assess the relationship between species abundance and environmental stressors. A difficult task is to interpret the relationships from such an analysis. We investigate the use of Bayesian Model Averaging as a tool to aid in the interpretation of relationships using canonical correspondence analysis. Illustrations on data indicate that the method allows investigation of model uncertainty and suggests relationships that may be missed using an ordinary analysis.
|