Anti Money Laundering Transaction Detection

Timeline: August 2024 - September 2024

• Project Description: A classifier that flags money laundering transactions in IBM’s synthetic AML dataset. Laundering transactions are pulled from all six dataset files and balanced into 886,163 rows. Features are built from time, account, currency and payment format data, and the project compares logistic regression, decision trees, random forest, extra trees, XGBoost, CatBoost, AdaBoost, an MLP and a Keras deep network. Built with Python, scikit-learn, XGBoost, CatBoost and TensorFlow on Kaggle.

• Project Github Link: money_laundering