Table 2 Optimised descriptor hyper-parameters for different data sets
From: Machine learning hydrogen adsorption on nanoclusters through structural descriptors
Data set | SOAP | ACSF | MBTR | ||
|---|---|---|---|---|---|
R c | R c | σ(k2) | σ(k3) | d | |
MoS2(single) | 6.0 | 5.0 | 0.078 | 0.075 | 0.3 |
Au40Cu40(single) | 6.0 | 6.0 | 0.078 | 0.075 | 0.3 |
MoS2(multi) | 8.0 | 8.0 | 0.015 | 0.05 | 0.3 |
AuCu(multi) | 8.0 | 10.0 | 0.015 | 0.05 | 0.7 |
