|
Activity Number:
|
359
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, August 2, 2016 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Statistics and the Environment
|
|
Abstract #318566
|
|
|
Title:
|
Modeling Nonstationarity in Space and Time with Dimension Expansion
|
|
Author(s):
|
Lyndsay Shand* and Bo Li
|
|
Companies:
|
and University of Illinois at Urbana-Champaign
|
|
Keywords:
|
Dimension expansion ;
Nonstationarity ;
Space-Time Random Field
|
|
Abstract:
|
We propose to model a spatio-temporal random field that has nonstationary covariance structure in both space and time domains by extending the dimension expansion method in Bornn et al. (2012). Simulations are conducted for both separable and nonseparable space-time covariance models, and the model is illustrated by Illinois wind speed and stream flow data. Both simulation and data analysis show that by modeling nonstationarity in both space and time improves the predictive performance over stationary covariance models or a model that is nonstationary in space but stationary in time.
|
Authors who are presenting talks have a * after their name.
