"What Did One Node Say To The Other? A Study Of Inference In Bayesian Networks"
J. Richard Drushal Jr 2007
Abstract
A Bayesian network is a graphical model for probabilistic relationships among a set of variables. Graphical models are a marriage between probability theory and graph theory. This thesis provides an understanding of Bayesian statistics, probability theory, and graph theory that is necessary to study Bayesian networks. Additionally, an overview of Bayesian networks, including construction, structural elements, applications and algorithms is provided.
The information contained in this thesis is sufficient for implementing an engine to analyze Bayesian networks. ISBayes is a software package created for this thesis to analyze Bayesian networks stored in a standard XML format. ISBayes implements the HUGIN inference algorithm, utilizing the Algebra of Bayesian Belief Universes. ISBayes is written in Java using the JGraphT and JGraph Application Programming Interfaces (API) for network modeling and display respectively.