|
Activity Number:
|
274
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, August 5, 2008 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Biometrics Section
|
| Abstract - #302624 |
|
Title:
|
Nugget Estimation for a Class of Nonparametric Semivariograms
|
|
Author(s):
|
Jeffrey Spence*+ and Patrick Carmack and Qihua Lin and Richard Gunst and William R. Schucany
|
|
Companies:
|
The University of Texas Southwestern Medical Center and The University of Texas Southwestern Medical Center and The University of Texas Southwestern Medical Center and Southern Methodist University and Southern Methodist University
|
|
Address:
|
5323 Harry Hines Blvd., Dallas, TX, 75390-8896,
|
|
Keywords:
|
brain imaging ; B spline ; fMRI ; kriging ; spatial modeling ; SPECT
|
|
Abstract:
|
Kriging estimators and their variances depend importantly on the behavior of the semivariogram near the origin. It is often the case that the semivariogram is discontinuous there when a nugget effect is present. Shapiro and Botha (1991) have offered a flexible class of nonparametric estimators based on a spectral representation of the covariance, but it necessarily extrapolates to the origin. In this paper we have incorporated an additive shift parameter in the nonlinear optimization routine used by Cherry et al. (1996) and estimated this feature by minimizing ISE with a spline-smoothed empirical estimate of the semivariogram. In simulation these estimates have significantly lower MSE compared to the parametric estimates from the same realizations. An application to functional brain imaging data is featured to demonstrate the automation advantages for multiple regional spatial modeling.
|
