Table 2.
Mean and standard deviation of the root-mean-squared error (RMSE) per patient over 100 different random initializations and the mean over all patients for predicting glucose levels 30 respectively 60 min into the future
| 30-min horizon | 60-min horizon | |||
|---|---|---|---|---|
| Patient ID | LSTM | t0 | LSTM | t0 |
| 559 | 18.773 ± 0.179 | 23.401 | 33.696 ± 0.365 | 39.404 |
| 570 | 15.959 ± 0.374 | 18.809 | 28.468 ± 0.834 | 31.577 |
| 588 | 18.538 ± 0.106 | 21.893 | 31.337 ± 0.210 | 35.928 |
| 563 | 17.961 ± 0.192 | 20.786 | 29.012 ± 0.169 | 34.032 |
| 575 | 21.675 ± 0.218 | 25.452 | 33.823 ± 0.268 | 39.164 |
| 591 | 20.294 ± 0.107 | 24.249 | 32.083 ± 0.182 | 38.219 |
| μ | 18.867 | 22.432 | 31.403 | 36.387 |
| σ | ± 1.794 | ± 2.217 | ± 2.078 | ± 2.860 |
t0 denotes the naive baseline of predicting the last value