Bayesian Networks



IDI offers a one, two, or three-day course in probabilistic inference and Bayesian Nets entitled Bayesian Network Analysis. The course provides a practical overview and introduction to the discipline of probabilistic inference. The practical, hands-on format introduces students to Bayesian Nets using the software package Netica. At the end of the three day Bayesian net course the student should: be aware of a spectrum of applications of Bayesian networks, be able to structure the nodes and arcs in a Bayesian network while working with experts, be familiar with Netica (a commercial software tool for Bayesian networks), and understand how to build Bayesian networks from data.

The first day of the course begins with a demonstration of the key features of Bayesian nets and a discussion of some of the diverse applications.  The basis of Bayesian nets in probability theory is then presented along with the algorithms used in Bayesian nets to perform probabilistic analysis. The first day closes with a demonstration of the Netica software package.

The second day of the course focuses on the formulation of real world problems as well as the “learning” feature of Bayesian networks. The day opens with a discussion of the standard model structures used in building Bayesian nets. Next, formulation of the nets is demonstrated using problems suggested by students. The resulting models are implemented and analyzed in Netica real time during the class. The third day of the course focuses on the “learning” feature of Bayesian nets.