Bank Fraud Detection ML Project

This project detects fraudulent bank accounts using machine learning. It leverages XGBoost, SMOTE for class imbalance, and includes preprocessing, evaluation, and visualization steps.

Features

• Data cleaning and preprocessing
• Feature scaling and encoding
• SMOTE for handling class imbalance
• XGBoost classifier
• ROC-AUC evaluation
• Modular Python code

Files

• main.py: Runs the ML pipeline and evaluates the model
• preprocessing.py: Data loading and preprocessing logic
• visualization.py: Fraud distribution plots
• requirements.txt: Project dependencies

Run Instructions

pip install -r requirements.txt
python main.py

Author

Habil Huseynov – M.S. in Applied Machine Learning @ University of Maryland