Bayesian belief networks are used to model application domains that are characterized by inherent uncertainty. The nodes of the network represent random variables, and the edges represent probabilistic dependencies between variables.

Speaking in terms of graph properties, Bayesian belief networks are directed acyclic graphs.

We suggest to represent Bayesian belief networks with upward drawings, i.e., drawings in which the eges tend to flow, as much as possible, in a common direction. Also, for Bayesian belief networks we suggest to represent edges with smoothed lines.

GAPI tutorial on upward drawings and BLAG tutorial on upward drawings provide a comprehensive guide on how to use GDT to construct these widely used drawings.

Last update : July 31, 2002
Website design by INTEGRA Sistemi,