Machine-Learning Prediction of the Computed Band Gaps of Double Perovskite Materials

19 May 2022, Version 1
This content is a preprint and has not undergone peer review at the time of posting.


Conventional electronic structure methods based on density functional theory (DFT) suffer from not only high computational cost that scales cubically with the number of electrons, but also limited accuracy arising from the approximations of the exchange-correlation functional. In particular, the latter issue is particularly severe for insulators and semiconductors, whose band gaps are systematically underestimated by DFT. Surrogate methods based on machine learning has garnered much attention as a viable alternative to bypass these limitations, especially in prediction of solid-state band gaps. Here, we construct a random forest regression model for band gaps of double perovskite materials, using a dataset of 1306 band gaps computed with the GLLBSC (Gritsenko, van Leeuwen, van Lenthe, and Baerends solid correlation) functional. Among the 20 physical features employed, we find that the bulk modulus, superconductivity temperature, and cation electronegativity exhibit the highest importance scores, consistent with the physics of the underlying electronic structure. Using the top 10 features, a model accuracy of 85.6% with a root mean square error of 0.64 eV is obtained, comparable to a previous study employing kernel ridge regression. Our results attest to the potential of machine learning regressions for rapid screening of promising candidate functional materials.


machine learning
density functional theory
computational materials science
electronic structure


Comments are not moderated before they are posted, but they can be removed by the site moderators if they are found to be in contravention of our Commenting Policy [opens in a new tab] - please read this policy before you post. Comments should be used for scholarly discussion of the content in question. You can find more information about how to use the commenting feature here [opens in a new tab] .
This site is protected by reCAPTCHA and the Google Privacy Policy [opens in a new tab] and Terms of Service [opens in a new tab] apply.