The Corresponding category about dataset "Coauthor CS"

shchur / gnn-benchmark

Framework for evaluating Graph Neural Network models on semi-supervised node classification task

https://arxiv.org/abs/1811.05868

MIT License

444 stars 73 forks source link

The Corresponding category about dataset "Coauthor CS" #6

Open JialongWang1224 opened 4 years ago

JialongWang1224 commented 4 years ago

Hello, author, First of all, thank you very much for your work! When we use the "Coauthor CS" dataset, we will get the class label as 0 to 9. And in your paper "Pitfalls of Graph Neural Network Evaluation" said that "and class labels indicate most active fields of study for each author."(page 7) Now I need the true class name the real "active fields" not the number. Can you give some help, thank you very much.😊

MengLiuPurdue commented 3 years ago

Hello, author,

I am also curious to know the true class name of Amazon dataset. Is there a way to get such info? Thanks so much!

shchur commented 3 years ago

You can get the original class names as follows

import numpy as np
d = dict(np.load("data/npz/amazon_electronics_computers.npz", allow_pickle=True))
print(d["class_names"])

Replace amazon_electronics_computers with the name of the dataset that you are interested in (see this folder for the list).

@JialongWang1224 I'm really sorry I didn't notice your original question.