This project focuses on building a fraud detection model for credit card transactions using a dataset containing transactions made by European cardholders in September 2013. With a highly unbalanced dataset, where frauds account for only 0.172% of all transactions, the challenge lies in effectively detecting fraudulent transactions while minimizing false positives. By employing machine learning techniques and possibly addressing class imbalance through methods like SMOTE, the aim is to develop a model that can accurately identify fraudulent transactions, thus preventing customers from being charged for unauthorized purchases.