Machine learning approaches for feature engineering of the crystal structure: Application to the prediction of the formation energy of cubic compounds

TL;DR

This paper introduces a scalable feature engineering framework for crystal structures, using 3D voxelization and principal component analysis, to improve prediction of formation energies in crystalline compounds.

Contribution

It presents a novel, efficient method for quantifying crystal structures and extracting low-dimensional features applicable to various surrogate models.

Findings

Top 25 features effectively predict formation energy

Framework works with Gaussian process and neural network models

Features improve surrogate model accuracy

Abstract

In this study, we present a novel approach along with the needed computational strategies for efficient and scalable feature engineering of the crystal structure in compounds of different chemical compositions. This approach utilizes a versatile and extensible framework for the quantification of a three-dimensional (3-D) voxelized crystal structure in the form of 2-point spatial correlations of multiple atomic attributes and performs principal component analysis to extract the low-dimensional features that could be used to build surrogate models for material properties of interest. An application of the proposed feature engineering framework is demonstrated on a case study involving the prediction of the formation energies of crystalline compounds using two vastly different surrogate model building strategies - local Gaussian process regression and neural networks. Specifically, it is shown that the top 25 features (i.e., principal component scores) identified by the proposed framework serve as good regressors for the formation energy of the crystalline substance for both model building strategies.