Activity Number:
|
121
- Handling Large Dimensionality, Skewness and Non-Stationarity Through Multi-Resolution Spatial Modeling
|
Type:
|
Topic Contributed
|
Date/Time:
|
Monday, July 29, 2019 : 8:30 AM to 10:20 AM
|
Sponsor:
|
Section on Statistics and the Environment
|
Abstract #303002
|
|
Title:
|
A Bi-Resolution Spatial Model Based on the Skew-T Distribution
|
Author(s):
|
Stefano Castruccio* and Felipe Tagle and Marc Genton
|
Companies:
|
University of Notre Dame and University of Notre Dame and King Abdullah University of Science and Technology
|
Keywords:
|
skew t;
wind power;
wind;
non Gaussian
|
Abstract:
|
The Gaussian assumption underlies much of the stochastic process theory for spatially-referenced datasets due to its analytical tractability. Several approaches have been developed in recent decades to reconcile this theory with non-Gaussian features that are commonly found in observational datasets across a broad range of applications. The skew-normal and skew-t distributions have only recently begun to appear in the spatial statistics literature. This article presents a spatial model based on the skew-t distribution that aids in modeling spatial configurations characterized by an interaction between a large-scale effect and region-specific fine-scale effects. The model's predictive ability is compared against another model in the literature, and illustrated using daily wind speeds over Saudi Arabia.
|
Authors who are presenting talks have a * after their name.