Table 2 Prediction of the streamflow peaks on the test set using data augmentation.
From: Artificial intelligence for streamflow prediction in river basins: a use case in Mar Menor
Horizon | Gauge station | Data augmentation technique | Best classification algorithm | Balanced accuracy |
|---|---|---|---|---|
1H | La Puebla | None | LR & RF | 0.5 |
IPIP | LR | 0.902 | ||
SMOTE | LR | 0.499 | ||
Resample | LR | 0.499 | ||
Desembocadura | None | LR | 0.625 | |
IPIP | LR | 0.734 | ||
SMOTE | LR | 0.75 | ||
Resample | LR | 0.75 | ||
12H | La Puebla | None | LR & RF | 0.5 |
IPIP | LR | 0.806 | ||
SMOTE | LR | 0.658 | ||
Resample | LR | 0.657 | ||
Desembocadura | None | LR & RF | 0.5 | |
IPIP | LR | 0.829 | ||
SMOTE | RF | 0.5 | ||
Resample | RF | 0.5 | ||
24H | La Puebla | None | LR | 0.5 |
IPIP | LR | 0.539 | ||
SMOTE | LR | 0.49 | ||
Resample | LR | 0.49 | ||
Desembocadura | None | LR & RF | 0.5 | |
IPIP | LR | 0.535 | ||
SMOTE | RF | 0.5 | ||
Resample | LR | 0.477 |
