|
Activity Number:
|
98
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Monday, August 7, 2006 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Statisticians in Defense and National Security
|
| Abstract - #305468 |
|
Title:
|
Cokriging with Generalized Cross-Covariances for Detecting Radioactivity
|
|
Author(s):
|
Chunfeng Huang*+ and Noel Cressie and Yonggang Yao and Tailen Hsing
|
|
Companies:
|
The Ohio State University and The Ohio State University and The Ohio State University and The Ohio State University
|
|
Address:
|
Cockins Hall, Room 404, Columbus, OH, 43210-1247,
|
|
Keywords:
|
cokriging ; generalized cross covariance ; spatial ; radioactivity ; sensors
|
|
Abstract:
|
Radioactivity detection and prediction are important for assessing the effect of a so-called 'dirty bomb' on people and their environment. Suppose there are two types of sensors deployed to augment each other---one being costly but providing high-quality information at a small number of key locations and the other being inexpensive with many sensors, low-quality information, and covering a larger area. If both types of observations are intrinsic random processes, we can introduce a generalized cross-covariance to describe their spatial cross-dependencies. A nonparametric method is proposed for estimation of all spatial dependencies. We can then use cokriging to predict radioactivity at unsampled locations based on both types of data.
|
