Table 1 Statistics of predicted bandgaps by ML algorithms for 4 × 4 and 5 × 5 supercell systems
From: Bandgap prediction by deep learning in configurationally hybridized graphene and boron nitride
MAE (eV) | MAEF | RMSE (eV) | RMSEF | R 2 | |
|---|---|---|---|---|---|
4 × 4 systems | |||||
VCN | 0.08 | 5.34% | 0.10 | 6.47% | 0.9483 |
RCN | 0.08 | 5.12% | 0.11 | 6.44% | 0.9426 |
CCN | 0.07 | 4.59% | 0.09 | 5.83% | 0.9547 |
SVM | 0.40 | 27.6% | 0.33 | 34.7% | 0.0029 |
5 × 5 systems | |||||
VCN | 0.07 | 5.72% | 0.09 | 7.57% | 0.9212 |
RCN | 0.08 | 6.39% | 0.10 | 8.33% | 0.9124 |
CCN | 0.15 | 13.3% | 0.16 | 15.7% | 0.9285 |
SVM | 0.27 | 22.8% | 0.51 | 29.5% | 0.0032 |
