The aim of the PRONEL project was to develop a prototype of a data mining tool that were able to extracting Bayesian network models from data in "collaboration" with an domain expert. This was achieved and a software prototype was developed. This prototype is available for download.
Pronel © is an European project ESPRIT
performed by four societies:
Siemens AG (Germany), Schlumberger (France),
Hugin (Denmark), and Tiga Technologies (France).
With the Pronel BN Learner tool you can load
a database table and where the columns represent variables/nodes
in the Bayesian network. Then, you can perform a structural
learning task that will give you a possibly incomplete Bayesian
network structure which may contain some uncertainty. Through
simple interactions with the user, the uncertainties are solved
and the final Bayesian network structure is determined. The last
part of the process consist of learning the underlying
probabilities of the Bayesian network.
