PKBoost-AI-Labs/PkBoost

PKBoost: Adaptive GBDT for Concept Drift, Built from scratch in Rust, PKBoost manages changing data distributions in fraud detection with a fraud rate of 0.2%. It shows less than 2% degradation under drift. In comparison, XGBoost experiences a 31.8% drop and LightGBM a 42.5% drop

40
/ 100
Emerging

PKBoost helps professionals building predictive models, especially for fraud or anomaly detection, by creating highly accurate models that automatically adapt to changing data patterns over time. You provide your historical data, and it outputs a robust predictive model capable of identifying rare events with high precision, even as trends shift. This is ideal for risk analysts, data scientists, and operations engineers who need reliable detection systems in dynamic environments.

Use this if you need to build a predictive model that must remain accurate in the face of evolving data, such as detecting new fraud patterns, identifying anomalies in real-time systems, or diagnosing diseases where indicators might subtly change.

Not ideal if your data distribution is completely static and unchanging, or if you primarily work with categorical features that would require extensive pre-processing.

fraud-detection anomaly-detection risk-assessment predictive-modeling real-time-monitoring
No Package No Dependents
Maintenance 10 / 25
Adoption 8 / 25
Maturity 13 / 25
Community 9 / 25

How are scores calculated?

Stars

66

Forks

5

Language

Rust

License

Apache-2.0

Last pushed

Jan 27, 2026

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/PKBoost-AI-Labs/PkBoost"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.