|
Activity Number:
|
587
- Recent Advances in Statistical Graphics
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 2, 2017 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Statistical Graphics
|
|
Abstract #323567
|
|
|
Title:
|
Dynamic Transelliptical Graphical Models for Sparse Precision Matrix Estimation
|
|
Author(s):
|
Tzu-Chun Wu* and Emily L. Kang
|
|
Companies:
|
University of Cincinnati and University of Cincinnati
|
|
Keywords:
|
Dynamic Structure ;
Graphical Model ;
High Dimensionality ;
Precision Matrix ;
Sparsity ;
Transelliptical Distribution
|
|
Abstract:
|
We propose a new method for estimating high dimensional precision matrices from dynamic transelliptical graphical (DTG) models. Also, we assume in the DTG model that the precision matrix is sparse and varies. This method combines not only the flexibility of kernel estimation to characterize the changing precision matrix, but also the extra robustness due to the transelliptical modeling. An efficient algorithm via linear programming is developed. The performance of DTG models is demonstrated through extensive simulation studies and applications to real data sets.
|
Authors who are presenting talks have a * after their name.
