|
Activity Number:
|
366
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, August 4, 2009 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Statistics and the Environment
|
| Abstract - #304421 |
|
Title:
|
Maximum Likelihood Estimation of Covariance Parameters in the Spatial Random Effects Model
|
|
Author(s):
|
Matthias Katzfuss*+ and Noel A. Cressie
|
|
Companies:
|
The Ohio State University and The Ohio State University
|
|
Address:
|
1958 Neil Ave, 404 Cockins Hall, Columbus, OH, 43210,
|
|
Keywords:
|
Covariance function ; Fixed rank kriging ; EM algorithm ; Geostatistics ; Remote sensing ; Spatial statistics
|
|
Abstract:
|
With the rapid increase in the size of geostatistical data sets, scalable methods for spatial data analysis, such as the spatial-random-effects model (Fixed Rank Kriging), are becoming increasingly important. We explore maximum likelihood estimation of the variance and covariance parameters in this model, and we propose the use of an E-M algorithm for this purpose. Some properties of the E-M estimators are discussed. We also compare their performance to that of the existing binned method-of-moments estimators on simulated data and on a global data set of model-based CO2.
|
