PASCAL - Pattern Analysis, Statistical Modelling and Computational Learning

A Lattice Approach to {B}ayesian Networks
Peter Harremoes
In: Conference on Uncertainty in Artificial Intelligence, 18-21/6 2009, Montreal, Canada.

Abstract

Bayesian networks and conditional independence is studied via functional dependences. Armstrong's axioms known from the theory of relational databases is used to reformulate the concept of Bayesian networks into the theory of join-semidistributive lattices. Lattice theory provides us with a richer language to discuss causation than graph theory does.

PDF - Requires Adobe Acrobat Reader or other PDF viewer.
EPrint Type:Conference or Workshop Item (Talk)
Project Keyword:Project Keyword UNSPECIFIED
Subjects:Computational, Information-Theoretic Learning with Statistics
ID Code:5070
Deposited By:Peter Harremoes
Deposited On:24 March 2009