Sigma
- Open-source conditional inference trees library in Python, designed for quick, robust and easy first-look diagnostic of data in production
- No black-box complexity - excellent robustness from rigorous statistical roots using statistical significance tests to select splits
- Simple API, minimal dependencies, production-ready from the start
Open Source
Sigma is freely available under an open-source license. Contributions, issues and feedback are welcome.
Coming soon...Key Features
Statistical Rigor
Conditional inference trees use permutation-based significance tests to select splits, avoiding the variable selection bias of traditional decision trees.
Production Ready
Lightweight and minimal dependencies, designed for quick diagnostics on real-world data in production environments.
Pure Python
Easy to install, easy to integrate, no complex build requirements. Works with standard data science tooling.