Page 72 - Read Online
P. 72
Tang et al. J. Mater. Inf. 2025, 5, 38 https://dx.doi.org/10.20517/jmi.2025.05 Page 13 of 19
Figure 7. Component functions of the additive model for Nb -Nb Si alloys. See Supplementary Figures 2-4 for the plots of the
I 5 3
component functions for other Nb Si alloys.
5 3
Figure 8. Component functions of the additive model for the combined dataset.
CONCLUSIONS
To elucidate fundamental alloying effects of multi-component alloys, it is crucial to develop accurate ML
methods to predict the formation energies of Nb and Nb-Nb Si eutectic alloys substituted with various
5
3
alloying elements in the Nb and Nb Si phases. Various complex factors influence the accuracy of ML
3
5
prediction. This case study is instructive, in particular, because it presents the problem of ML of materials
properties from relatively sparse data often encountered in doped materials systems, and it presents a typical
doped system where the data can be naturally divided into subsets based on the type of substitutional site.
