Auto-CheML: Auto Chemical Machine Learning

Auto-CheML is an integrated automated modeling software designed for clustering analysis, feature engineering, and energy predictions. Its workflow comprises four main procedures: analyzing data distribution, clustering the filtered data into distinct classes, selecting important features through data correlation or mutual information analysis, and optimizing machine learning parameters, including hyperparameter tuning, performance evaluation, and model interpretation. Throughout the entire process, Auto-CheML systematically documents all generated data in organized tables with automatically generated legends, ensuring data clarity and facilitating reproducibility. Users only need to provide a data table and a JSON configuration file, which can be easily created using the tool’s built-in Graphical User Interface (GUI). Auto-CheML has been successfully applied to CO₂ reduction reaction (CO₂RR) and nitrogen reduction reaction (NRR) datasets, demonstrating its effectiveness in supporting in-depth research and analysis. To request the datasets, please register and complete the LICENSE FORM, scan and email it to majing@nju.edu.cn.