Skip to main content
. 2019 Jan 29;6(9):1801367. doi: 10.1002/advs.201801367

Table 1.

Summary of the RMSE for the 16 excitations and the RSE for spectra for the 6k and the 132k datasets. The results are averages over 5 runs, except for the spectra predictions of 132k dataset which were averaged over 3 runs. The resulting statistical error is at most ±0.003 and has therefore been omitted from this table

Datasets → 6k 132k
Model ↓ Levels [eV] Spectra Levels [eV] Spectra
MLP 0.317 NA NA NA
CNN 0.304 0.057 0.231 0.039
DTNN 0.251 0.051 0.186 0.029