🔄 Transactions Fraud Detection API

A Flask-based API that predicts fraudulent online payment transactions using a machine learning model trained on real transaction data.

---

🌐 Hosted API

---

📌 Overview

This API accepts JSON-formatted transaction details and returns a fraud prediction based on a trained ML model.
It is intended to be used by external applications (e.g., an ASP.NET MVC).

---

🧠 Machine Learning Model

• Algorithm: (e.g., Random Forest / XGBoost)
• Training Notebook: Online-Payment-Fraud-Detection.ipynb
• Model Input: Cleaned and one-hot encoded transaction data
• Output:
  • 0 → Legitimate transaction
  • 1 → Fraudulent transaction

---

🔧 API Usage

POST /predict

URL:
https://transactionsmlapi-production.up.railway.app/predict

Request Headers:

Content-Type: application/json

📥 Request Format

Send a JSON object with the following structure (sample):

{
  "type": "CASH_OUT",
  "amount": 9800,
  "oldbalanceOrg": 9800,
  "newbalanceOrig": 0,
  "oldbalanceDest": 0,
  "newbalanceDest": 0
}

📥 Response Format

{
  "prediction": "Fraud"  // or "Not Fraud"
}