💳 Credit Card Fraud Detection using Neural Networks

This project focuses on detecting fraudulent transactions using a Multi-Layer Perceptron (MLP) neural network. It tackles the challenge of highly imbalanced datasets with oversampling techniques and evaluates model performance using key classification metrics.

---

📌 Project Overview

• Developed a neural network model (MLP) to detect fraud in credit card transactions.
• Addressed severe class imbalance using ADASYN (Adaptive Synthetic Sampling).
• Trained and compared two MLP models:
  • A basic MLP architecture.
  • An MLP with Dropout to reduce overfitting.
• Performed hyperparameter tuning using GridSearchCV.
• Implemented Early Stopping to prevent overtraining.
• Evaluated model performance using:
  • Accuracy
  • F1-Score
  • ROC-AUC Score

---

🔧 Technologies & Libraries

• Python
• NumPy, Pandas
• Scikit-learn
• Keras / TensorFlow
• ADASYN (from imblearn)
• Matplotlib, Seaborn (for visualization)

---

🚀 Results

• The dropout-enhanced MLP showed improved generalization and better fraud detection performance.
• Achieved higher F1-score and ROC-AUC compared to the basic model.
• Model training was more efficient due to early stopping and optimized hyperparameters.

---

🧠 Key Concepts

• Imbalanced Learning
• Synthetic Oversampling (ADASYN)
• Dropout Regularization
• GridSearchCV for Tuning
• Early Stopping in Training
• Binary Classification Metrics

---