|
Activity Number:
|
586
- Frontiers of Multivariate Spatial Methodology
|
|
Type:
|
Invited
|
|
Date/Time:
|
Thursday, August 1, 2019 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Statistics and the Environment
|
|
Abstract #300085
|
|
|
Title:
|
Pushing the Limits of Multivariate Spatial Models: How Many Is Too Many?
|
|
Author(s):
|
William Kleiber* and Mitchell Krock and Dorit Hammerling
|
|
Companies:
|
University of Colorado and University of Colorado at Boulder and National Center for Atmospheric Research
|
|
Keywords:
|
basis function;
multivariate;
climate model;
numerical weather prediction
|
|
Abstract:
|
Multivariate spatial and space-time datasets are ubiquitous in many sciences, and only in the past two decades has the statistical community begun developing and exploring models for such processes. Most extant models have been applied only to datasets with a handful of variables with much work focused on the bivariate problem. In this talk we review some challenges for multivariate modeling, and discuss the relationship between model parsimony and utility. We propose a new framework to model hundreds of variables simultaneously motivated by data from the Community Atmosphere Model (CAM).
|
Authors who are presenting talks have a * after their name.
