Credit Card Fraud Detection

This project focuses on detecting fraudulent credit card transactions using advanced anomaly detection techniques in machine learning and deep learning. Given the highly imbalanced nature of fraud data, we implemented models such as Isolation Forest, Autoencoder, One-Class Neural Network (OC-NN), and Deep Support Vector Data Description (Deep SVDD) to identify suspicious activity based on deviations from normal patterns. The workflow included data preprocessing, exploratory data analysis, model training on legitimate transactions only, and evaluation using metrics like Precision, Recall, and F1-Score to handle the class imbalance. The resulting models successfully flagged fraudulent transactions with high accuracy while minimizing false positives, offering a robust solution for enhancing financial security.