|
Activity Number:
|
112
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, July 30, 2007 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Statistics and the Environment
|
| Abstract - #309974 |
|
Title:
|
Assessing Statistical Spatial Performance: An Application in Archaeology
|
|
Author(s):
|
Yuemei Wang*+ and Lance Waller and Zev Ross
|
|
Companies:
|
Emory University and Emory University and ZevRoss Spatial Analysis
|
|
Address:
|
Rollins School of Public Health, Atlanta, GA, 30322,
|
|
Keywords:
|
statistical spatial performance ; Generalized Additive Models ; receiver operating characteristic (ROC) curve
|
|
Abstract:
|
We investigate the problem of cluster detection when adjusting for covariates using Generalized Additive Models (GAM). GAMs are extensions of generalized linear models replacing some or all parametric terms with smooth functions. In addition, GAMs provide spatial smoothing for 2-dimensional data by using thin-plate spline regression. We apply GAMs to data from the Black Mesa archaeological project to identify clusters of early versus late Anasazi settlement sites while adjusting for exposure to rivers around those sites. We also evaluate the statistical power of the approach as a function of space. We generate the receiver operating characteristic (ROC) curve associated with classifications of locations based on Monte Carlo simulations and compare the GAM results with results based on kernel density estimation of the early-to-late relative risk surface and SaTscan cluster statistics.
|
