A combined first-principles and machine learning study of phase stability and thermoelectric properties in ZrSiPt

stratified cross-validation; successfully classifies the crystal phase (∼90% accuracy) and predicts key thermoelectric properties. Feature-importance analysis identifies lattice constant and band gap as the dominant descriptors, directly linking geometric packing and electronic stability to phase selection. This combined DFT+ML approach not only elucidates the dual-phase behavior of ZrSiPt but also demonstrates a synergistic workflow for accelerating the design of phase-specific materials for thermoelectric, optoelectronic, and mechanically robust applications.