This repository contains the code implementation of the paper "Improved topology features for node classification on heterophilic graphs". To reproduce the results in the paper, you need to follow these steps:

1. Download the datasets from https://github.com/CUAI/Non-Homophily-Large-Scale.git and https://github.com/RecklessRonan/GloGNN.git
2. Use

mkdir save_embd
python main.py --dataset snap-patents --directed --method gcnt --adj_order 3 --tebd_type ours --tebd_dim 512 --treduc_save 

to precompute the needed feature.

3.Use

python main.py --dataset snap-patents --directed  --device cuda:1 --adj_order 1  --tebd_dim 512 --hidden_channels 32  --method gat_pc  --w1 4 --tdropout 0.5 --dropout 0.5 --input_dropout 0.5  --tebd_type ours --runs 1 --epochs 300 --display_step 25 --gat_heads 4

to reproduce the reported result.

If you find our work useful, please cite our paper by

@InProceedings{10.1007/978-3-031-70368-3_7,
author="Lai, Yurui
and Zhang, Taiyan
and Fan, Rui",
editor="Bifet, Albert
and Davis, Jesse
and Krilavi{\v{c}}ius, Tomas
and Kull, Meelis
and Ntoutsi, Eirini
and {\v{Z}}liobait{\.{e}}, Indr{\.{e}}",
title="Improved Topology Features for Node Classification on Heterophilic Graphs",
booktitle="Machine Learning and Knowledge Discovery in Databases. Research Track",
year="2024",
publisher="Springer Nature Switzerland",
address="Cham",
pages="105--123",
isbn="978-3-031-70368-3"
}