PCTs updates

[ben_mrad]

Considering that we have a Bayesian Network already learned. At time t new data has been collected. Can we update the PCTs with just the new data, i.e. without redoing the parameters learning with a new dataset containing the old and the new data.

[marek [BayesFusion]]

Absolutely! Start with "Learning parameters" from the "Learning" menu. In the "Learn Parameters with EM" dialog, select "Keep original" (parameters) and give the existing data a confidence level that corresponds to the number of records that the existing parameters are based on. GeNIe will weight the new records (it knows obviously the number of the new records) against the records that the existing parameters are based on. This is described of course in GeNIe manual.
I hope this helps,

Marek