user-risk-detection-api

FastAPI-based microservice that predicts the risk level of a user session

Overview • Key Features • Risk Classes • Model Performance • Dataset Generation • API Example • Usage • Download • Emailware • Contributing • License • Contact • Acknowledgments

Overview

user-risk-detection-api is a FastAPI-based microservice that predicts the risk level of a user session using behavioral signals such as:

• IP changes
• device switching
• location anomalies
• login time patterns
• session duration

The model outputs a risk classification (0 / 1 / 2) along with probabilities.

Key Features

• FastAPI REST API with /docs (Swagger UI)
• ML model powered by XGBoost
• Multiclass classification (safe / medium / risky)
• Probability output for each class
• Trained on large synthetic dataset (500k+ rows)
• Deterministic + noise-injected data generation
• Fraud-oriented feature engineering
• CORS enabled (frontend ready)

Risk Classes

Label	Meaning
0	Safe
1	Medium Risk
2	High Risk

Model Performance

• Accuracy: ~94–96%
• Balanced multiclass performance
• Robust against noisy inputs
• Tuned fraud detection threshold

Dataset Generation

Synthetic dataset includes:

• 500,000+ unique rows
• Zero duplicates
• Controlled class balance
• Realistic noise injection

API Example

Endpoint

POST /api/v1/predict
Content-Type: application/json

Request

{
  "session_duration": 180,
  "is_revoked": 0,
  "ip_change_count": 2,
  "device_change_count": 1,
  "location_change": 0,
  "login_hour": 10,
  "is_night_login": 0,
  "os_variation": 2
}

Response

{
  "prediction": 1,
  "confidence": 0.87,
  "probabilities": [0.1, 0.87, 0.03]
}

Usage

git clone https://github.com/zouari-oss/user-risk-detection-api
cd user-risk-detection-api/project

chmod +x setup && ./setup
fastapi run

Note

API available at: http://localhost:8000/docs

Download

You can download the latest installable version of user-risk-detection-api for Windows, macOS and Linux.

Emailware

user-risk-detection-api is an emailware. Meaning, if you liked using this app or it has helped you in any way, would like you send as an email at zouariomar20@gmail.com about anything you'd want to say about this software. I'd really appreciate it!

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

License

This repository is licensed under the GPL-3.0 License. You are free to use, modify, and distribute the content. See the LICENSE file for details.

Contact

For questions or suggestions, feel free to reach out:

• GitHub: zouari-oss
• Email: zouariomar20@gmail.com
• LinkedIn: zouari-omar

Acknowledgments

Built with ❤️ for the open-source community.