Integration of PAW-Based DFT Data with Machine Learning for the Prediction of 2D Magnetic Material Candidates
Predicting magnetic ordering in 2D materials from chemical composition. C2DB and V2DB merged into 141,171 compositions, gradient boosting, ROC-AUC 0.986 and 90.6% recall on the magnetic class. Poster award at the department's thesis presentations (January 2026).