Hi, I just read the original paper on ProNE and wanted to test it on the CTU-13 dataset which consist of network communication data between IPs and I planned to use ProNE to embedd the communication graph which, ideally would be a weighted graph, with edge weights proportional to the amount of messages between each IP.
So here is my question: is ProNE supposed to work with a weighted graph? If so, does your implementation support it? (or if not, would it be enough to change the matrix0 definition in proNE.py line 32/33, assigning the corresponding weight instead of 1 in the adjacency matrix?)
Hi, I just read the original paper on ProNE and wanted to test it on the CTU-13 dataset which consist of network communication data between IPs and I planned to use ProNE to embedd the communication graph which, ideally would be a weighted graph, with edge weights proportional to the amount of messages between each IP.
So here is my question: is ProNE supposed to work with a weighted graph? If so, does your implementation support it? (or if not, would it be enough to change the matrix0 definition in proNE.py line 32/33, assigning the corresponding weight instead of 1 in the adjacency matrix?)