|
Activity Number:
|
59
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Sunday, August 3, 2008 : 4:00 PM to 5:50 PM
|
|
Sponsor:
|
Section on Statistics and the Environment
|
| Abstract - #300454 |
|
Title:
|
Infill Asymptotic Properties of Tapered Maximum Likelihood Estimators
|
|
Author(s):
|
Juan Du*+ and Hao Zhang and V. S. Mandrekar
|
|
Companies:
|
Michigan State University and Purdue University and Michigan State University
|
|
Address:
|
Department of Statistics and Probability, MSU, East Lansing, MI, 48824-1027,
|
|
Keywords:
|
Covariance tapering ; infill asymptotics ; Matern covariance functions ; maximum likelihood estimator
|
|
Abstract:
|
Massive spatial and spatio-temporal data are common these days, and the analysis of these massive data involves operations of a large covariance matrix. Repeatedly, inversion of the covariance matrix is needed in both the maximum likelihood and Bayesian inferences. When the sample size is extremely large, operations on the large covariance matrix are a numerical challenge. Covariance tapering is a technique to alleviate the numerical challenges. We investigate how the tapering affects the asymptotic efficiency of the maximum likelihood estimator (MLE) and establish the asymptotic properties, particularly the asymptotic distribution of the exact MLE and tapered MLE under the infill asymptotic framework for Matern model. We show that the tapered MLE is asymptotically as efficient as the true MLE for the Matern model under some tail conditions on the tapering function.
|
