Activity Number:
|
442
- Model-Based Statistical Analysis of Network Data
|
Type:
|
Invited
|
Date/Time:
|
Wednesday, August 2, 2017 : 8:30 AM to 10:20 AM
|
Sponsor:
|
IMS
|
Abstract #322353
|
View Presentation
|
Title:
|
Random Networks, Graphical Models, and Exchangeability
|
Author(s):
|
Alessandro Rinaldo* and Steffen Lilholt Lauritzen and Kayvan Sadeghi
|
Companies:
|
Carnegie Mellon University and University of Copenhagen and University of Cambridge
|
Keywords:
|
Network Models ;
Graphical Models ;
Exchangeability
|
Abstract:
|
We investigate connections among exponential random network models (ERGM), bidirected and undirected graphical models, and exchangeability in network models. We show that exchangeable finite network models can be well approximated by mixtures of curved exponential families corresponding to a distinguished class of graphical models for marginal independence of binary data. We obtain a simple derivation of de-Finetti theorem for exchangeable arrays, and we link it to the theory of graphons. Using this characterization, we discuss the challenges and intrinsic difficulties of fitting exchangeable network models. This is joint work with S. Lauritzen and K. Sadeghi.
|
Authors who are presenting talks have a * after their name.
Back to the full JSM 2017 program
|