|
Activity Number:
|
282
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Tuesday, August 5, 2003 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Bayesian Stat. Sciences
|
| Abstract - #300565 |
|
Title:
|
Bayesian Multiscale Modeling
|
|
Author(s):
|
Marco Antonio Rosa Ferreira*+
|
|
Companies:
|
Universidade Federal Do Rio De Janeiro
|
|
Address:
|
R. Barao De Mesquita 380/205, Rio De Janerio, RJ 20540-003, , , Brazil
|
|
Keywords:
|
Bayesian analysis ; random fields ; physical sciences ; multiscale modeling ; computer models
|
|
Abstract:
|
We introduce two classes of multiscale models: One for time series and another for random fields. The novel framework couples standard Markov models for the time series or the random field at different levels of aggregation, and links the levels via stochastic links to induce new and rich classes of structured linear models. Our construction has several interesting characteristics: with just a few parameters, it produces a great variety of autocorrelation functions for time series and variograms for random fields; it has the ability to coherently and efficiently combine information from different scales; and it has can emulate long memory processes for time series and for random fields. Several examples of multiscale time series analysis are presented as well as an example of the application of multiscale random fields to the estimation of permeability fields in subsurface hydrology studies. In the particular case of permeability estimation, there are available data of different nature and at different levels of resolution. Our multiscale framework allows the incorporation of all the available information in a coherent way.
|
