Networks and number of reticulations

[larruskito]

Hi all,

I am trying to get some networks through the funtion InferNetwork_ML. The software works perfectly and the results are great, but is there any methodology to assign the best number of reticulations. The values of Pseudolikelihood are always better and better adding new reticulations (more complexity to de model?), but, how can I penalize this increase the number of reticulation in orden to obtain a proper model?

Thanks for your help