EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Gradient boosted and statistical feature selection workflow for materials property predictions

Son Gyo Jung iD, Guwon Jung iD, Jacqueline M Cole iD

The Journal of Chemical Physics · 21.11.2023 · 159 (19)

DOI: 10.1063/5.0171540 ↗

Kurz gefasst

Fachartikel in der Zeitschrift The Journal of Chemical Physics, online erschienen am 16.11.2023. Verfasst von Son Gyo Jung und 2 weiteren Autor:innen. Beteiligte Einrichtung: Cavendish Laboratory, Department of Physics, University of Cambridge, J. J. Thomson Avenue, Cambridge CB3 0HE, United Kingdom. Die Arbeit wurde bisher 30-mal zitiert (Crossref) und nennt 128 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Royal Academy of Engineering, ISIS Neutron and Muon Source, Rutherford Appleton Laboratory.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

With the emergence of big data initiatives and the wealth of available chemical data, data-driven approaches are becoming a vital component of materials discovery pipelines or workflows. The screening of materials using machine-learning models, in particular, is increasingly gaining momentum to accelerate the discovery of new materials. However, the black-box treatment of machine-learning methods suffers from a lack of model interpretability, as feature relevance and interactions can be overlooked or disregarded. In addition, naive approaches to model training often lead to irrelevant features being used which necessitates the need for various regularization techniques to achieve model generalization; this incurs a high computational cost. We present a feature-selection workflow that overcomes this problem by leveraging a gradient boosting framework and statistical feature analyses to identify a subset of features, in a recursive manner, which maximizes their relevance to the target variable or classes. We subsequently obtain minimal feature redundancy through multicollinearity reduction by performing feature correlation and hierarchical cluster analyses. The features are further refined using a wrapper method, which follows a greedy search approach by evaluating all possible feature combinations against the evaluation criterion. A case study on elastic material-property prediction and a case study on the classification of materials by their metallicity are used to illustrate the use of our proposed workflow; although it is highly general, as demonstrated through our wider subsequent prediction of various material properties. Our Bayesian-optimized machine-learning models generated results, without the use of regularization techniques, which are comparable to the state-of-the-art that are reported in the scientific literature.

Autor:innen und Institutionen

  • Cavendish Laboratory, Department of Physics, University of Cambridge, J. J. Thomson Avenue, Cambridge CB3 0HE, United Kingdom.
  • ISIS Neutron and Muon Source, STFC Rutherford Appleton Laboratory, Harwell Science and Innovation Campus, Didcot, Oxfordshire OX11 0QX, United Kingdom.
  • Research Complex at Harwell, Rutherford Appleton Laboratory, Harwell Science and Innovation Campus, Didcot, Oxfordshire OX11 0FA, United Kingdom.
  • Scientific Computing Department, STFC Rutherford Appleton Laboratory, Harwell Science and Innovation Campus, Didcot, Oxfordshire OX11 0QX, United Kingdom.

ORCID-Profile: Son Gyo Jung, Guwon Jung, Jacqueline M Cole

Förderung

  • Royal Academy of Engineering
  • ISIS Neutron and Muon Source
  • Rutherford Appleton Laboratory

Zitieren

APA

Jung, S. G., Jung, G., & Cole, J. M. (2023). Gradient boosted and statistical feature selection workflow for materials property predictions. The Journal of Chemical Physics, 159(19). https://doi.org/10.1063/5.0171540

Vancouver

Jung SG, Jung G, Cole JM. Gradient boosted and statistical feature selection workflow for materials property predictions. The Journal of Chemical Physics. 2023;159(19). doi:10.1063/5.0171540

BibTeX
@article{Jung2023,
  author = {Jung, Son Gyo and Jung, Guwon and Cole, Jacqueline M},
  title = {Gradient boosted and statistical feature selection workflow for materials property predictions},
  journal = {The Journal of Chemical Physics},
  year = {2023},
  volume = {159},
  number = {19},
  doi = {10.1063/5.0171540}
}
Infokarte | EUVIMED

Artikeldaten derzeit nicht verfügbar.

Recherche