How important is microstuctural feature selection for data-driven structure-property mapping?

Published

January 6, 2022

Author(s)

Hao Liu, Berkay Yucel, Daniel Wheeler, Baskar Ganapathysubramanian, Surya Kalidindi, Olga Wodo

Abstract

Data-driven approaches now allow for systematic mapping of microstructure to properties. In particular, we now have d\ iverse approaches to "featurize" microstructures, creating a large pool of machine-readable descriptors for subsequent stru\ cture-property analysis. We explore three questions in this work: (a) Can a small subset of features be selected to train a\ good structure-property predictive model? (b) Is this subset agnostic to the choice of feature selection algorithm? And (c\ ) can the addition of expert-identified features improve model performance? Using a canonical dataset, we answer in the aff\ irmative for all three questions.

Citation

MRS Communications

Volume

12

Issue

1

Pub Type

Journals

Download Paper

https://doi.org/10.1557/s43579-021-00147-4

Local Download

Keywords

materials-science, feature-selection, machine-learning, structure-property

Citation

Liu, H. , Yucel, B. , Wheeler, D. , Ganapathysubramanian, B. , Kalidindi, S. and Wodo, O. (2022), How important is microstuctural feature selection for data-driven structure-property mapping?, MRS Communications, [online], https://doi.org/10.1557/s43579-021-00147-4, https://tsapps.nist.gov/publication/get_pdf.cfm?pub_id=932830 (Accessed October 8, 2025)