TABLE IV.
Mean normalized absolute difference (MNAD) for different coarse-scale PDEs. “GP” and “NN” represent “Gaussian processes” and “Neural networks,” respectively.
| MNADu | MNADv | |
|---|---|---|
| No feature selection with GPs | 1.59 × 10−02 | 1.62 × 10−02 |
| No feature selection with NNs | 1.53 × 10−02 | 1.56 × 10−02 |
| Feature selection 1 with GPs | 1.58 × 10−02 | 1.62 × 10−02 |
| Feature selection 1 with NNs | 1.54 × 10−02 | 1.57 × 10−02 |
| Feature selection 2 with GPs | 2.39 × 10−02 | 2.20 × 10−02 |
| Feature selection 2 with NNs | 2.00 × 10−02 | 2.11 × 10−02 |
| Feature selection 3 with GPs | 3.20 × 10−02 | 3.31 × 10−02 |
| Feature selection 3 with NNs | 2.08 × 10−02 | 2.16 × 10−02 |