Revisiting Self-Supervised Heterogeneous Graph Learning from Spectral Clustering Perspective

This repository contains the reference code for the manuscript ``Revisiting Self-Supervised Heterogeneous Graph Learning from Spectral Clustering Perspective"

Contents

0. Installation
1. Preparation
2. Training

Installation

• pip install -r requirements.txt
• Unzip the datasets (datasets can be found in ./dataset/)

Preparation

Important args:

• --use_pretrain Test checkpoints to reproduce the results
• --dataset Heterogeneous graph: ACM, Yelp, DBLP, Aminer
• --custom_key Node: node classification

Training

python main.py