💳 Credit Card Fraud Detection System

This project uses machine learning to detect fraudulent credit card transactions. The dataset used is highly imbalanced and contains only a small fraction of fraudulent transactions. This system is designed to help financial institutions identify suspicious activity effectively and minimize potential financial losses.

📌 Project Overview

• Goal: Accurately classify credit card transactions as fraudulent or legitimate.
• Dataset: European credit card transactions (from Kaggle).
• Tech Stack: Python, Pandas, Scikit-Learn, Matplotlib, Seaborn

---

🚀 Features

• Data preprocessing and cleaning
• Exploratory Data Analysis (EDA) and visualization
• Feature selection and scaling
• Model training using multiple ML algorithms
• Evaluation using accuracy, precision, recall, F1-score, and ROC-AUC
• Handling imbalanced data using SMOTE and undersampling techniques
• Confusion matrix and ROC curve visualization

---

📂 Folder Structure

Credit_Card_Fraud_Detection_System/
│
├── data/                        # Dataset files (CSV)
├── models/                      # Saved model files
├── notebooks/                   # Jupyter notebooks
├── visuals/                     # Plots and graphs
├── fraud_detection.py          # Main Python script
├── requirements.txt            # Dependencies
└── README.md                   # Project documentation

---

📈 Evaluation Metrics

• Confusion Matrix
• Accuracy
• Precision
• Recall
• F1 Score
• ROC-AUC Curve

---

📊 Algorithms Used

• Logistic Regression
• Random Forest Classifier
• Decision Trees
• XGBoost
• Support Vector Machine (SVM)
• K-Nearest Neighbors (KNN)

---

📚 Future Improvements

• Integrate deep learning (LSTM or autoencoders)
• Real-time fraud detection with stream data
• Web interface or dashboard using Flask/Streamlit

🤝 Acknowledgements

• Kaggle - Credit Card Fraud Detection Dataset
• Scikit-learn, Matplotlib, Seaborn, Imbalanced-learn