HOME > MAST-ML

MAST-ML

MAST-ML (Materials Simulation Toolkit for Machine Learning) is an open-source Python toolkit focused on supervised learning workflows for materials research. It provides a configurable pipeline for data preprocessing, feature generation/selection, model training, and evaluation, with tutorials and examples to accelerate end‑to‑end studies.

Information
Case / Article
Learning Content
Review
Developer’s Voice

Information

Official site	https://github.com/uw-cmg/MAST-ML
Openness	★★★
License	MIT License
Core Developers	Developers: University of Wisconsin–Madison Computational Materials Group, led by Prof. Dane Morgan, with contributors listed in the project README and documentation.
Related Papers	Jacobs et al., “The Materials Simulation Toolkit for Machine Learning (MAST‑ML): An automated open source toolkit to accelerate data‑driven materials research,” Computational Materials Science 175 (2020), 109544. Palmer et al., “Calibration after bootstrap for accurate uncertainty quantification in regression models,” npj Computational Materials 8, 115 (2022). Schultz et al., “A general approach for determining applicability domain of machine learning models,” npj Computational Materials 11, 95 (2025).
Other	Description: MAST-ML emphasizes configurable, reproducible supervised-learning studies. It supports scikit‑learn model workflows, integrates feature engineering and selection, and provides documentation/tutorials for data import, model comparison, error analysis, uncertainty quantification, and domain‑of‑applicability checks.