Weighted networks?

[anthonimes]

Hello!

I am successfully using Walklets so far, so this is more a question than an issue. :)
Is it possible to adapt your code to work with edge-weighted networks?

It is mentioned in your article that this can be done, but being a stranger to this domain I am afraid to do things the wrong way.
The only attempt I made so far was to change this part of the code:

for edge in tqdm(self.G.edges(data=True)):
    self.G[edge[0]][edge[1]]['weight'] = edge[2]['weight'] # 1.0
    self.G[edge[1]][edge[0]]['weight'] = edge[2]['weight'] # 1.0

in file walkers.py. But this actually does not seem to make any difference.

I apologize if my question is irrelevant or has a really easy answer,
Thanks in advance for your help!
Anthony

[benedekrozemberczki]

Hi there Anthony,

You have to write a sampler for random sampling of neighbors - ideally you would use a binary search tree or hopscotch sampling.

It is not a trivial problem.

Benedek