Table 5. Root-mean-square errors, mean absolute errors and maximum absolute errors for the residual fields presented in Fig. 5 ▸ .
The result of the Gaussian process regression is compared with the weighted least-squares fit (WLSQ). Values are unitless (strain) and on the same scale (10−4) as in Fig. 5 ▸.
| RMSE | MAE | Maximum absolute error | ||||
|---|---|---|---|---|---|---|
| Strain | GP | WLSQ | GP | WLSQ | GP | WLSQ |
| εxx | 1.322 | 2.076 | 1.101 | 1.737 | 2.791 | 7.42 |
| εyy | 1.042 | 1.371 | 0.846 | 0.999 | 3.856 | 6.094 |
| εzz | 0.887 | 1.489 | 0.769 | 1.157 | 1.914 | 6.736 |
| εxy | 1.122 | 1.511 | 0.955 | 1.172 | 2.778 | 6.306 |
| εxz | 0.24 | 1.04 | 0.198 | 0.798 | 1.506 | 4.85 |
| εyz | 0.48 | 0.958 | 0.399 | 0.742 | 1.34 | 4.652 |