Credit Card Fraud Detection Using Deep Learning

Credit Card Fraud Detection Using Deep Learning

Team Project — A collaborative AI project focused on detecting fraudulent credit card transactions using deep learning techniques.

The project includes:

- Data cleaning and preprocessing

- Handling class imbalance using SMOTE

- Feature scaling and preparation

- Exploratory data analysis and visualization

- Training a Multi-Layer Perceptron (MLP)

- Building an Autoencoder for fraud detection

- Selecting model thresholds using validation data

- Combining the outputs of multiple deep learning models

- Evaluating the final system using Precision, Recall, F1-Score, ROC-AUC, and PR-AUC

- Generating the final Fraud / Not Fraud prediction

The project demonstrates an end-to-end AI workflow for an imbalanced fraud detection problem, combining supervised and unsupervised deep learning approaches as part of a collaborative team effort.