Yes, except the point was to ensure a DAG for any given sparse graph, not just any old (in this case a maximally dense) graph on those variables. As it happens, it’s indeed trivial or perhaps axiomatic to Bayes nets that there’s a DAG.
Yes, except the point was to ensure a DAG for any given sparse graph, not just any old (in this case a maximally dense) graph on those variables. As it happens, it’s indeed trivial or perhaps axiomatic to Bayes nets that there’s a DAG.