Ensemble-machine-learning-based correlation analysis of internal and band characteristics of thermoelectric materials

Machine learning can significantly help to predict the thermoelectric properties of materials, such as the Seebeck coefficient and electrical conductivity. However, the mechanism underlying the excellent performance of such models is not known. In this study, a new dual-route machine learning system (DMLS) is developed to extract the relationship between the features from materials and the ones from band structure. These findings can help us to set up a bridge between the feature significance and the thermal electric properties, such as Seebeck coefficient, which can provide theoretical guidance regarding the designing of a material with excellent thermoelectric properties.