AutoDetect: Credit Card Fraud Detection

A deep learning–based system designed to detect fraudulent credit card transactions using an autoencoder. The model was trained exclusively on normal transactions to identify anomalies with high accuracy on highly imbalanced datasets.

Key Features:

Achieved 99.62% accuracy on fraud detection.

Built using one-class autoencoder for anomaly detection.

Optimized model performance using the Detection Error Tradeoff (DET) curve.

Designed to handle real-world, imbalanced financial data.