QPGCN: Graph Convolutional Network with a Quadratic Polynomial Filter for Overcoming Over-smoothings

This repository is the official implementation of QPGCN: Graph Convolutional Network with a Quadratic Polynomial Filter for Overcoming Over-smoothing.

Requirements

To install requirements:

pip install -r requirements.txt

Training

To train the model(s) in the paper, run this command:

python run.py --config ./config/QPGCN-cora.json
python run.py --config ./config/QPGCN-citeseer.json
python run.py --config ./config/QPGCN-pubmed.json
python run.py --config ./config/QPGCN-dblp.json

Results

Our model achieves the following performance on :

Model name	Cora	Citeseer	Pubmed	DBLP
QPGCN	83.31 $\pm$ 0.30%	71.22 $\pm$ 0.44%	79.22 $\pm$ 0.40%	85.92 $\pm$ 0.72%