🛡️ Credit Card Fraud Detection API

A production-ready, containerized FastAPI service that predicts fraudulent credit card transactions using a trained machine learning model (e.g., XGBoost, Logistic Regression, etc.).

---

📊 Dataset

• Dataset: Credit Card Fraud Detection
• Number of records: 284,807 transactions
• Class distribution: Highly imbalanced (Class 1 ≈ 0.17%)

---

🧠 Model Performance (XGBoost)

Validation Report:

Metric	Class 0 (Non-Fraud)	Class 1 (Fraud)
Precision	0.9997	0.8737
Recall	0.9998	0.8469
F1-Score	0.9998	0.8601

Accuracy: 99.95%
Confusion Matrix:

              | Predicted Not Fraud     | Predicted Fraud
Not Fraud     | 56852                   | 12
Fraud         | 15                      | 83

🧠 Project Features

• ✅ Pretrained model selection based on highest F1-score
• ✅ Input validation with Pydantic
• ✅ Preprocessing pipeline before prediction
• ✅ Dockerized for deployment
• ✅ Testing with pytest
• ✅ Real API testing using requests

🚀 Running the Project

Clone the repository

git clone https://github.com/MahmoodAbdali79/fraud-detection-api.git
cd fraud-detection-api

Run locally

pip install -r requirements.txt
python -m web.main

Run with Docker

docker build -t fraud-api .
docker run -p 8000:8000 fraud-api

Access API

curl -X 'POST' \
  'http://0.0.0.0:8080/predict' \
  -H 'accept: application/json' \
  -H 'Content-Type: application/json' \
  -d '{
  "features":
{"Time":53937.0,
"V1":-2.0426081994,
"V2":1.5735780737,
"V3":-2.3726524327,
...
}

the input data is not complete. find complete inpute data there.