Table 3 Summary of materials property prediction tools.
Property prediction | Model | Training dataset | Performance | Output |
|---|---|---|---|---|
2D materials | Random Forest | 2DMatPedia Material Project | 88.98% (Acc) | Label |
| Â | Â | Â | Â | Probability |
Noncentro symmetry | Random Forest | Material Project | 84.8% (Acc) | Label |
| Â | Â | Â | Â | Probability |
Band gap | Roost DeeperGATGNN | Material Project | 0.465 (MAE) | Band gap |
| Â | Â | Â | Â | (eV) |
Elastic moduli | CrabNet DeeperGATGNN | 12858 samples from MP | 15.7 (MAE, Bulk) | Bulk mod (GPa) |
| Â | Â | Â | 18 (MAE, Shear) | Shear mod (GPa) |
|  |  |  | 76.8 (MAE, Young’s) | Young’s mod (psi) |
| Â | Â | Â | 8.7 (MAE, Poisson) | Poisson ratio |
Hardness | Roost DeeperGATGNN | 12854 samples from MP | 2.42 (MAE) | Hardness (N/mm2) |
Thermal conductivity | CrabNet DeeperGATGNN | 2688 samples from ICSD | 5.03 (MAE) | Thermal conductivity (W/(mK)) |
Ionic conductivity | under development | N/A | N/A | Ionic conductivity |
Superconductivity | Random Forest CrabNet | 25378 samples from supercon | 4.76 (MAE) | Transition temperature (K) |
