Hey,
I have been using genie and smile for a while and I am looking for a solution to learn the parameters of a bayesian network from a data set with soft observations.
For example:
We have modelled a bayesian networks with 3 nodes (theft, earthquake and alarm). We can observe alarm with hard evidence, saying there was an alarm or there was no alarm. We are not sure if there was an earthquake, so we can say that there was an earthquake with 80% belief (soft evidence). With these observations we want to calculate how likely theft is.
Now we want to learn the parameters of this networks using a data set with soft and hard evidence.
Do you know to a possibility to do this in Genie (or in smilearn)
Thank you
Simon
Learning BN parameters from soft data set
Re: Learning BN parameters from soft data set
Learning from soft evidence is not supported in GeNIe and SMILE.