Data and Supplemental information for predicting the thermodynamic stability of perovskite oxides using machine learning models

Publication date: August 2018
Source:Data in Brief, Volume 19
Author(s): Wei Li, Ryan Jacobs, Dane Morgan
To better present the machine learning work and the data used, we prepared a supplemental spreadsheet to organize the full training dataset prepared from DFT calculations, the individual elemental properties, the generated element-based descriptors derived from the elements present in each perovskite composition, and lists of the specific features selected and used our machine learning models. We have also provided supplemental information which contains additional details related to our machine learning models which were not provided in the main text (Li et al., In press) [1]. In particular, the supplemental information provides results on training and testing five regression models (using the same data and descriptors as the regression of Ehull in main text) to predict the formation energies of perovskite oxides. We provided source code that trains the machine learning models on the provided training dataset and predicts the stability for the test data.



https://ift.tt/2J56vQw