This is the code implementation of the paper "Enhancing Fraud Detection in GNNs with Synthetic Fraud Generation and Integrated Structural Features" (ICANN 2024).

Dataset

YelpChi and Amazon can be downloaded from here.

Run inside the data directory wget https://github.com/YingtongDou/CARE-GNN/raw/master/data/YelpChi.zip && unzip YelpChi.zip

Run python src/data_process.py to pre-process the data.

Usage

python main.py --config ./config/brie_yelpchi.yml

Brief Description

Introduction

Contribution - BRIE

Additional structural features (Contribution 1)

Synthesizing fraud nodes (Contribution 2)

Results

Conclusion