leo-q-316/imgagn

Imbalanced Network Embedding vi aGenerative Adversarial Graph Networks

imgagn's Issues

IndexError: index 2 is out of bounds for axis 0 with size 2

Hello, this work looks like so great!
But it is reported that "IndexError: index 2 is out of bounds for axis 0 with size 2" when running train.py.

About the code of DR-GCN

Hello, I haven't found the code of DR-GCN. If there is, could you share it with me? Thank you very much.

Dataset PubMed

Hey! Thanks for sharing your souce code, but can you provide the dataset PubMed you use in the paper? The standard class distribution of PubMed is {1: 7739, 0: 4103, 2: 7875}, so if I reconstruct it as an imbalanced binary dataset with the smallest class as the minority class (as described in your paper), the ratio of minority class will be 24.9% instead of 5.25%. So can you provide the PubMed dataset you use?

Many thanks.

code & paper problem

Hi，this is a very good work, but why can only 0/1 binary classification problems be processed in utils? Isn't his ultimate goal to perform better classification?

The whole dataset

Hello, congratulations to this interesting work!
But there is something wrong with the code when loading the dataset.
Could you please share the entire "Cora" data set you are using?

Dataset

Hi,
Thank you for sharing this awesome work.
But can you put the dataset up?