Pick and Choose: A GNN-based Imbalanced Learning Approach for Fraud Detection

Benchmark Model Rank Results
fraud-detection-on-amazon-fraudPC-GNN#4AUC-ROC: 95.86Averaged Precision: 85.49
fraud-detection-on-yelp-fraudPC-GNN#8AUC-ROC: 79.87Averaged Precision: 48.10
node-classification-on-amazon-fraudPC-GNN#4AUC-ROC: 95.86
node-classification-on-yelpchiPC-GNN#7AUC-ROC: 79.87