Premium
Crystal structure representations for machine learning models of formation energies
Author(s) -
Faber Felix,
Lindmaa Alexander,
von Lilienfeld O. Anatole,
Armiento Rickard
Publication year - 2015
Publication title -
international journal of quantum chemistry
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.484
H-Index - 105
eISSN - 1097-461X
pISSN - 0020-7608
DOI - 10.1002/qua.24917
Subject(s) - ansatz , coulomb , crystal (programming language) , matrix (chemical analysis) , crystal structure , unit sphere , molecule , physics , statistical physics , mathematics , chemistry , pure mathematics , crystallography , quantum mechanics , computer science , electron , chromatography , programming language
We introduce and evaluate a set of feature vector representations of crystal structures for machine learning (ML) models of formation energies of solids. ML models of atomization energies of organic molecules have been successful using a Coulomb matrix representation of the molecule. We consider three ways to generalize such representations to periodic systems: (i) a matrix where each element is related to the Ewald sum of the electrostatic interaction between two different atoms in the unit cell repeated over the lattice; (ii) an extended Coulomb‐like matrix that takes into account a number of neighboring unit cells; and (iii) an ansatz that mimics the periodicity and the basic features of the elements in the Ewald sum matrix using a sine function of the crystal coordinates of the atoms. The representations are compared for a Laplacian kernel with Manhattan norm, trained to reproduce formation energies using a dataset of 3938 crystal structures obtained from the Materials Project. For training sets consisting of 3000 crystals, the generalization error in predicting formation energies of new structures corresponds to (i) 0.49, (ii) 0.64, and (iii) 0.37 eV / atom for the respective representations. © 2015 Wiley Periodicals, Inc.