Sustainable Thermoelectric Materials Predicted by Machine Learning

Abstract Using datasets from several sources, a list of more than 450 materials is generated and related them with their thermoelectric properties. This obtained by generating set features using only the molecular formula. Subsequently, machine learning algorithm classifies in specific, binary classes, for example, possessing high or low Seebeck coefficients electrical conductivity. After adjusting threshold values grouping into clusters, performance 25k predicted. Finally, results are filtered to obtain sustainable materials, that is, neither toxic nor critical, (ideally) inexpensive, isotropic regard transport properties simplify preparation procedure.