|
Activity Number:
|
353
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Tuesday, August 4, 2009 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Nonparametric Statistics
|
| Abstract - #305144 |
|
Title:
|
Hierarchical Bayesian Models for Predicting Spatially Correlated Curves
|
|
Author(s):
|
Joon Jin Song*+ and Bani K. Mallick
|
|
Companies:
|
University of Arkansas and Texas A&M University
|
|
Address:
|
Department of Mathematical Sciences, Fayetteville, AR, 72701,
|
|
Keywords:
|
Functional Data Analysis ; Hierarchical Bayesian Model ; Spatial Process ; Wavelet
|
|
Abstract:
|
Functional data analysis has emerged as a new area of statistical research with wide range of application. Functional data typically consists of curves that are ordered measurement on some interval. We propose some novel models based on wavelets for spatially correlated functional data. The proposed models enable one to regularize curves observed over space as to predict curves at unobserved sites. We have compared the performance of these Bayesian models with several priors on the wavelet coefficients using the posterior predictive criterion. The proposed models are employed to analyze a real porosity data set in petroleum engineering.
|
