Hello!
I have recently started using the bnlearn package to create Bayesian Networks from data. My dataset has missing values so I'm using the structural.em command. I have found that as I increase the number of iterations (using max.iter), the number of edges/arcs in the network also increases. Is that how it is supposed to be? I have tried Bayesian Networks elsewhere too but I haven't seen any relation between number of iterations and number of edges, except in R.
Thank you!