Fig. 2: Important features for the regression and classification ML models. | npj Computational Materials

Fig. 2: Important features for the regression and classification ML models.

From: Accelerated design and discovery of perovskites with high conductivity for energy applications through machine learning

Fig. 2

a The conductivity of the 7230 different perovskites considered from literature with respect to the most important features from the XG-Boost regression model showing higher conductivities for higher minimum electronegativity of B-site ions and lower average B-site ion size. b The type of charge carrier (proton, oxygen, mixed protonic/electronic, mixed oxide/electronic, mixed protonic/oxide, and electronic conductors) with respect to the most important features from the XG-Boost classifier model, showing that the lower B size atomic mass and higher minimum formation energy of B-site oxides regions are mostly electronic or mixed oxide/electronic conductors. All the proton-conducting perovskites (which are of interest for low temperature conductivities) have relatively high B-site atomic mass or lower formation energy of B-site oxide. The range of conductivities for each charge carrier is shown in the inset.

Back to article page