Table 2.

Summary of the ML model results in terms of correlation coefficient (R²) for both the training and hold out testing set.

	Gaussian process	Neural network	Gradient boosting machine
9–12% Cr FMA
Training set	0.93 ± 0.05	0.94 ± 0.01	0.99 ± 2 × 10−4
Testing set	0.92 ± 0.07	0.93 ± 0.02	0.98 ± 4 × 10−3
Austenitic stainless steel
Training set	0.91 ± 0.06	0.86 ± 0.02	0.99 ± 3 × 10−4
Testing set	0.83 ± 0.08	0.84 ± 0.02	0.95 ± 5 × 10−3