|
Activity Number:
|
109
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, August 7, 2006 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Statistics and the Environment
|
| Abstract - #306619 |
|
Title:
|
Search for Multivariate Structure for EMAP Fish Data Using Partition Modeling Approach
|
|
Author(s):
|
Feng Gao*+ and Eric P. Smith and Samantha C. Prins
|
|
Companies:
|
Virginia Polytechnic Institute and State University and Virginia Polytechnic Institute and State University and Virginia Polytechnic Institute and State University
|
|
Address:
|
1218 University City Blvd., Blacksburg, VA, 24060,
|
|
Keywords:
|
Voronoi tessellation ; partition model ; fish data ; RDA/CCA
|
|
Abstract:
|
A multivariate partition modeling approach - a method of clustering sites with multivariate abundance of species as response in order to find the stressor-response relationships of interest will be presented. The method uses random Voronoi tessellations assign sites to one of k clusters that subdivide a region. The BIC-like optimal criterion and hot-spot detection criterion are proposed to find the 'best' clustering or cluster of interest. The BIC-like criterion uses proportion of constrained inertia variance) over total unconstrained inertia (variance) in Canonical Correspondence Analysis (CCA)/Redundancy Analysis (RDA) as a R-square measurement. The method then is applied to EMAP fish data to search for underlying multivariate structure.
|
