Sample graph size
gebaltso opened this issue · 2 comments
Hello and congrats on your great job. I m currently working on local community detection and I want to use your approach too. I want to ask about the sample graph size. I have a graph of 1000nodes and I want to make experiments with the whole graph and with sample graph. The problem is that I cannot use any size for the sample graph as I get a KeyError in b = map_dict[j] of sample_graph. Should I change any other parameter besides new_graph_size ? Hope you understand my question and thanks in advance.
I met the same problem.
I met the same problem.
Hello there!
To run it without sampling (all nodes) change the sample_rate to 1.
The problem I mentioned a month ago, was in the loops of def sample_graph:
"for i in range(new_graph_size):
for j in range(new_graph_size)"
i and j runs with unitary increase. I noticed that above those loops the map_dict{} is created but has some numbers missing. So there are cases of missing keys. The strange is that this happens not all the times but in specific seed nodes given. Any help welcome.