Classification of ABO3 perovskite solids: a machine learning study. Issue 5 (1st October 2015)

• Record Type: Journal Article
• Authors: Pilania, G.; Balachandran, P. V.; Gubernatis, J. E.; Lookman, T.
• Is Part Of: Acta crystallographica. Volume 71:Issue 5(2015:Oct.)
• Page Start: 507
• Page End: 513
• Publication Date: 2015-10-01
• DOI: 10.1107/S2052520615013979
• Languages: English
• ISSNs: 2052-5206

Abstract

<abstract abstract-type="main" xml:lang="en"> <title> <x xml:space="preserve">Abstract</x> </title> <p>We explored the use of machine learning methods for classifying whether a particular <italic>AB</italic>O₃ chemistry forms a perovskite or non‐perovskite structured solid. Starting with three sets of feature pairs (the tolerance and octahedral factors, the <italic>A</italic> and <italic>B</italic> ionic radii relative to the radius of O, and the bond valence distances between the <italic>A</italic> and <italic>B</italic> ions from the O atoms), we used machine learning to create a hyper‐dimensional partial dependency structure plot using all three feature pairs or any two of them. Doing so increased the accuracy of our predictions by 2–3 percentage points over using any one pair. We also included the Mendeleev numbers of the <italic>A</italic> and <italic>B</italic> atoms to this set of feature pairs. Doing this and using the capabilities of our machine learning algorithm, the gradient tree boosting classifier, enabled us to generate a new type of structure plot that has the simplicity of one based on using just the Mendeleev numbers, but with the added advantages of having a higher accuracy and providing a measure of likelihood of the predicted structure.</p> </abstract>