Real-Time Fraud Detection System

A machine learning-based fraud detection system that analyzes transaction patterns in real-time using behavioral analytics and anomaly detection.

🚀 Quick Start

For detailed step-by-step instructions, see QUICK_START.md

Using Docker (Recommended - Easiest)

# 1. Start all services
docker-compose up

# 2. In a new terminal, train ML models (first time only)
docker-compose exec backend python -m app.models.train_models

# 3. Access the application
# Frontend: http://localhost:5173
# Backend API: http://localhost:8000/docs

Manual Setup

See QUICK_START.md for complete manual setup instructions.

Tech Stack

Backend & ML

• Python 3.11+
• FastAPI - Modern async API framework
• Scikit-learn - Isolation Forest for anomaly detection
• XGBoost - Behavioral pattern classification
• Redis - Real-time data storage and caching
• PostgreSQL - Transaction history and audit logs

Frontend

• React 18+ with TypeScript
• Vite - Fast build tool
• Recharts - Real-time visualizations
• Tailwind CSS - Modern UI styling
• WebSocket - Real-time updates

Features

• Real-Time Detection: Instant analysis of transactions as they occur
• Behavioral Analytics: Learns individual user patterns to reduce false positives
• Anomaly Detection: Identifies both known and novel fraud patterns
• Live Dashboard: Real-time visualization of transactions and fraud alerts
• Risk Scoring: 0-100 risk score for each transaction

Documentation

• SETUP.md - Detailed setup instructions

Project Structure

fraud-detection-system/
├── backend/           # FastAPI backend and ML models
├── frontend/          # React dashboard
├── docker-compose.yml # Container orchestration
└── README.md