GCN_GAT

GCN and GAT

---

• authors GCN : Thomas N. Kipf , Max Welling
• authors GAT: Petar Velickovic, Guillem Cucurull, Arantxa Casanova, Arantxa Casanova, Pietro Lio, Yoshua Bengio
• link: https://arxiv.org/abs/1609.02907
• link: https://arxiv.org/abs/1710.10903
• file structure:

└── cresci-2015,Twibot-20,Twibot-22
    ├── main.py  # train model on the processed data
    ├── dataset.py  # preprocess data
    ├── utils.py  
    └── model.py  # load GCN/GAT model 

How to reproduce:

1. specify the dataset by running dataset=Twibot-22 in Dataset.py (Twibot-22 for example) ;

2. change the model in the model.py

3. train model by running:

   python main.py

Result:

random seed: 100, 200, 300, 400, 500

GCN

dataset		acc	precison	recall	f1
Cresci-2015	mean	0.9637	0.9559	0.9881	0.9717
Cresci-2015	std	0.0057	0.0069	0.0020	0.0043
Twibot-20	mean	0.7753	0.7523	0.8762	0.8086
Twibot-20	std	0.0173	0.0308	0.0331	0.0068
Twibot-22	mean	0.7839	0.7119	0.4480	0.5496
Twibot-22	std	0.0009	0.0128	0.0171	0.0091

GAT

dataset		acc	precison	recall	f1
Cresci-2015	mean	0.9689	0.9610	0.9911	0.9758
Cresci-2015	std	0.0021	0.0071	0.0051	0.0015
Twibot-20	mean	0.8327	0.8139	0.8953	0.8525
Twibot-20	std	0.0056	0.0118	0.0087	0.0038
Twibot-22	mean	0.7948	0.7623	0.4412	0.5586
Twibot-22	std	0.0009	0.0139	0.0165	0.0101

baseline	acc on Twibot-22	f1 on Twibot-22	type	tags
Kipf et al.	0.7839	0.5496	F T G	Graph Neural Network
Velickovic et al.	0.7948	0.5586	F T G	Graph Neural Network