Table 2 Comparison of test set prediction performance between models using augmented and original data

From: Data-augmented machine learning design and performance-enhancing quaternary synergistic mechanism of novel Cu-Be alloy

Model

R2

Predicted errors (%)

Augmented data (strength)

0.96

6

Original data (strength)

0.95

6

Augmented data (conductivity)

0.90

8

Original data (conductivity)

0.88

11

Augmented data (stress relaxation rate)

0.86

9

Original data (stress relaxation rate)

0.20

39