Closed dgoldsb closed 5 years ago
The first generator can be pretty suboptimal, I should first just try to get some results.
For the first generator I am just going to pick the inputs and outputs of the rules uniformly, as well as the number of links. Especially when we are dealing with motifs of 3 to 4 genes there is not much of a point in generating Barabasi-Albert networks (yet).
I need to reconsider the generator a bit, to make sure it does not create things that are not networks. This includes looking at the possible rules again.
I improved this function: I consider all possible edges, and select a rule for a sample without replacement from these edges.