Table 1 Summary of the evaluation metrics of the results obtained in the validation set, the independent test set and through the 10-fold cross-validation (CV). Data are reported as median ± median absolute deviation
From: Deriving three one dimensional NMR spectra from a single experiment through machine learning
MRE% ± MAD | RMSE ± MAD | R2 ± MAD | RPD ± MAD | |
|---|---|---|---|---|
Validation set (n = 321) | ||||
CPMG | 4.52 ± 0.916 | 3.24·10–4 ± 8.96·10–5 | 0.998 ± 0.00106 | 21.4 ± 6.40 |
Diffusion-edited | 2.00 ± 0.579 | 2.76·10–4 ± 9.63·10–5 | 1.00 ± 0.000291 | 40.9 ± 16.7 |
pJRES (from predicted CPMG) | 8.80 ± 1.59 | 1.40·10–2 ± 3.61·10–3 | 0.994 ± 0.00291 | 11.4 ± 3.05 |
pJRES (from NOESY) | 8.83 ± 1.49 | 1.36·10–2 ± 3.47·10–3 | 0.994 ± 0.00266 | 11.5 ± 2.98 |
Independent test set (n = 232) | ||||
CPMG | 5.97 ± 1.26 | 5.47·10–4 ± 1.35·10–4 | 0.995 ± 0.00205 | 12.7 ± 2.90 |
Diffusion-edited | 3.80 ± 1.19 | 7.85·10–4 ± 3.56·10–4 | 0.998 ± 0.00106 | 17.7 ± 6.31 |
pJRES (from predicted CPMG) | 12.6 ± 1.33 | 2.22·10–2 ± 5.01·10–3 | 0.968 ± 0.0162 | 5.03 ± 1.28 |
pJRES (from NOESY) | 13.3 ± 1.81 | 2.60·10–2 ± 5.48·10–3 | 0.960 ± 0.0192 | 4.25 ± 1.04 |
10-fold CV on the entire dataset (n = 1753) | ||||
CPMG | 4.50 ± 0.962 | 3.96·10–4 ± 1.16·10–4 | 0.997 ± 0.00139 | 18.7 ± 5.70 |
Diffusion-edited | 2.50 ± 0.857 | 3.67·10–4 ± 1.46·10–4 | 0.999 ± 0.000434 | 33.0 ± 13.0 |
pJRES (from predicted CPMG) | 9.59 ± 2.00 | 1.55·10–2 ± 4.34·10–3 | 0.992 ± 0.00465 | 9.87 ± 2.98 |
pJRES (from NOESY) | 9.00 ± 1.82 | 1.47·10–2 ± 4.20·10–3 | 0.993 ± 0.00412 | 10.5 ± 3.24 |
