|
Activity Number:
|
509
- New Approaches to Modeling and Inference for Complex Space-Time Data
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Wednesday, August 1, 2018 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Physical and Engineering Sciences
|
|
Abstract #330632
|
Presentation
|
|
Title:
|
Identification of Management Zone Using a Spatial Clustering Time-Varying Lattice Models
|
|
Author(s):
|
Youngdeok Hwang* and Huijing Jiang and Rodrigue Ngueyep
|
|
Companies:
|
Sungkyunkwan University and IBM Research and IBM Research
|
|
Keywords:
|
clustering;
lattice model;
lattice model;
time-varying coefficients
|
|
Abstract:
|
When making a policy decision, it is important to identify the regions in which the policy would lead to a desirable outcome. When both covariates and responses are collected across geographical regions over time, such impacts are often spatially clustered and time-varying. We model this time-varying nature via a spatial lattice model with spatio-temporal varying coefficients. We propose a penalized likelihood method to incorporate the spatial structure while efficiently estimating the model parameters. We demostrate the merits of our method via simulation studies and a real life application.
|
Authors who are presenting talks have a * after their name.
