Run python3 btc_training.py
in CARE-GNN folder. Using the default parameters and running for 250 epochs, we achieve:
Training Loss: 0.26482430371371185
f1: 0.6208869814020028
Precision: 0.6420118343195266
Recall: 0.6011080332409973
AUC: 0.8723118520826214