We are using the Elliptic Data Set (https://www.kaggle.com/ellipticco/elliptic-data-set) and working to improve on the orignals results by Weber, Mark, et al. "Anti-money laundering in bitcoin: Experimenting with graph convolutional networks for financial forensics." arXiv preprint arXiv:1908.02591 (2019).
We use ML models such as SVM and Random Forest. Oversampling techiques such as SMOTE, ctCAN and GMM were used to address the class imbalance problem in the dataset. Also, we generated node embeddings using DeepWalk and trained ML models on them.
Team Members - Bhavay Aggarwal, Saad Ahmad, Prasham Narayan