|
Activity Number:
|
317
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, August 5, 2008 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Statistical Computing
|
| Abstract - #301762 |
|
Title:
|
Collapsibility for Directed Acyclic Graphs
|
|
Author(s):
|
Xianchao Xie*+ and Zhi Geng
|
|
Companies:
|
Harvard University and Peking University
|
|
Address:
|
Department of Statistics , Cambridge, MA, 02215,
|
|
Keywords:
|
Graphical model ; directed acyclic graph ; collapsibility ; model collapsibility ; marginalization
|
|
Abstract:
|
Collapsibility means that certain aspects of a model are preserved after marginalization over some variables. In this talk, we present conditions for three kinds of collapsibility for directed acyclic graphs (DAGs): estimate collapsibility, conditional independence collapsibility and model collapsibility. We show that unlike in graphical loglinear models or hierarchical loglinear models, the estimate collapsibility and the model collapsibility are not equivalent in DAG models. We discuss the relationship among them and illustrate how the results obtained can be applied in simplifying the inference problems in DAGs. Algorithms are also given to find a minimum variable set containing a subset of variables of interest onto which a DAG model is collapsible.
|
