Table A1.
Performance metrics for neural network on validation set: The parameters of the meal detection and estimation framework are selected based on performance of the model on the validation set. This is based on last 20% of the training data in order to attain the predictive accuracy and prediction interval necessary.
| ID | Metric | ||
|---|---|---|---|
| RMSE | MAE | PI | |
| 1 | 5.8 | 4.0 | 95.3 |
| 2 | 5.9 | 4.1 | 94.5 |
| 3 | 8.0 | 4.6 | 94.7 |
| 4 | 5.7 | 4.0 | 94.9 |
| 5 | 7.7 | 4.3 | 93.6 |
| 6 | 5.7 | 4.1 | 94.9 |
| 7 | 6.7 | 4.8 | 94.7 |
| 8 | 7.2 | 4.3 | 93.7 |
| 9 | 5.0 | 3.7 | 97.3 |
| 10 | 8.4 | 4.3 | 95.4 |
| Average | 6.6 ± 1.1 | 4.2 ± 0.3 | 94.9 ± 1.0 |