Bayesian Networks

A Bayesian Network is a directed acyclic graph, where the nodes of the graph represent random variables or uncertain quantities and the links or arcs between the nodes represent the causal relationships between variables. The strength of the links is quantified by the nodes conditional probability table. Bayes theorem is used to resolve uncertainties in the Network. They were first describe by Judea Pearl in his book ....

I have used Bayesian Networks as the primary tool to model Sydney Harbour, as there graphical structure allows the experts that I dealing with to easily understand what I am representing in the model. Bayesian Networks have also allowed me to take advantage of the causal knowledge that the experts have in the domain. For the creation of Bayesian Networks I have used both the Netica graphical user interface (gui) and  API.

Here is a copy of the Chapter (Chapter.ps) of my thesis on Bayesian Networks for more detail.

See the Netica website - www.norsys.com for more information about their software here