🚀 Credit Card Fraud Detection

This project develops a machine learning model to detect fraudulent credit card transactions using the Kaggle Credit Card Fraud Detection dataset. It focuses on handling class imbalance, feature engineering, and optimizing model performance.

---

📂 Dataset

• Source: Kaggle - Credit Card Fraud Detection
• Size: 284,807 transactions
• Fraud Cases: ~0.17% (highly imbalanced)

---

🏗️ Approach

✅ Data Preprocessing

✔ Removed missing/duplicate values
✔ Scaled transaction amounts
✔ Created new features (e.g., transaction frequency, spending patterns)

✅ Handling Imbalanced Data

✔ Used SMOTE (Synthetic Minority Oversampling Technique) to balance fraud cases

✅ Model Training & Evaluation

✔ Models tested: Logistic Regression, Random Forest, XGBoost
✔ Used AUC-ROC, Precision-Recall, and Confusion Matrix for evaluation

---

📊 Model Performance

• Random Forest AUC-ROC: ~0.95
• XGBoost AUC-ROC: ~0.98
• False Positive Rate (FPR) reduced significantly with feature engineering

---

🛠️ Installation & Usage

1️⃣ Clone Repository

git clone https://github.com/yourusername/Credit-Card-Fraud-Detection.git
cd Credit-Card-Fraud-Detection

2️⃣ Install Dependencies

pip install -r requirements.txt