A descriptor combining eight elemental properties with pairwise distance-weighted interaction matrices improves LightGBM predictions of FM/FiM ordering (82.4% accuracy) and magnetic moment (CC 0.94) across 5,741 stable binary and ternary compounds.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cond-mat.mtrl-sci 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Advancing Magnetic Materials Discovery -- A structure-based machine learning approach for magnetic ordering and magnetic moment prediction
A descriptor combining eight elemental properties with pairwise distance-weighted interaction matrices improves LightGBM predictions of FM/FiM ordering (82.4% accuracy) and magnetic moment (CC 0.94) across 5,741 stable binary and ternary compounds.