:: Workshop – Bayesianism II: Foundations and Applications ::

Invited Talk 10 (friday – 14h00)
Bayesian Models On Event Trees And The Chain Event Graph.
Jim Q Smith (The University of Warwick).

Abstract:
Chain Event Graphs (CEGs) provide a framework for modeling many asymmetric discrete problems graphically. They arise naturally from descriptions about how situations can unfold. They are also a much more general class of models than the discrete Bayesian Network - which is a simple special case of a CEG. Conditional independence interrogation, conjugate estimation and fast model selection procedures are all possible within this new class. Finally they provide a framework for expressing and exploring causal hypotheses in a much more flexible way than is possible using a Causal Bayesian Network. I will use a number of simple examples to illustrate the various ways this new class of graphical models can be used effectively.

Promotion and Support:
			CAPES-Proex Pós–graduação em Estatística-USP