How to perform structure learning for Bayesian network given already partially constructed Bayesian network?

Question

Let's assume that we have a dataset of variables (random events)I apriori would like to set dependency conditions between some of them and perform structure learning to figure out the rest of the Bayesian network.

How can this be done practically (e.g. some libraries, like bnlearn) or, at least, in theory?

I was trying to google it, but haven't found anything related.

Answer 1

The hill climbing or the constraint-based structure learning algorithms accept whitelist or blacklist arguments permitting or prohibitting some arcs.