Machine learning (ML) is increasingly becoming a helpful tool in the searchfor novel functional compounds. Here we use classification via random foreststo predict the stability of half-Heusler (HH) compounds, using onlyexperimentally reported compounds as a training set. Cross-validation yields anexcellent agreement between the fraction of compounds classified as stable andthe actual fraction of truly stable compounds in the ICSD. The ML model is thenemployed to screen 71,178 different 1:1:1 compositions, yielding 481 likelystable candidates. The predicted stability of HH compounds from three previoushigh throughput ab initio studies is critically analyzed from the perspectiveof the alternative ML approach. The incomplete consistency among the threeseparate ab initio studies and between them and the ML predictions suggeststhat additional factors beyond those considered by ab initio phase stabilitycalculations might be determinant to the stability of the compounds. Suchfactors can include configurational entropies and quasiharmonic contributions.
展开▼