How can I reproduce the experiment using only the adjacency matrix without node features?

[Faith0918]

Hi, I'm a huge fan of yours.

I have an issue reproducing your experiments using only the adjacency matrix without node features.

I would like to ask you how should it be coded to reproduce the performance in your paper trained without node features.

If it's okay with you, I would love to learn from your code.

Thank you.

[xlleng]

Hey guys
maybe I could answer your question

You can use this code:

import numpy as np
import networkx as nx
features = np.identity(adj.shape[0])
features = nx.adjacency_matrix(nx.from_numpy_matrix(features))

Make the feature matrix as a indentity matrix
wish this would help you

[Faith0918]

Thank you for your reply. I appreciate it:)) Best regards. Shinsil Kang

…

On Mon, Apr 19, 2021, 10:09 PM xlleng ***@***.***> wrote: Hey guys maybe I could answer your question You can you this code: import numpy as np import networkx as nx features = np.identity(adj.shape[0]) features = nx.adjacency_matrix(nx.from_numpy_matrix(features)) Make the feature matrix as a indentity matrix wish this would help you — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#71 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AM2RDLPNMVMEHZXGETAFBPDTJQTSHANCNFSM42E6LPMQ> .