Table 3 Experiments La Puebla.
From: Artificial intelligence for streamflow prediction in river basins: a use case in Mar Menor
Horizon | Algorithm | MAE | RMSE | CVRMSE | WI | NSE |
|---|---|---|---|---|---|---|
1H | DT | 0.013 | 0.134 | 206.498 | 0.804 | \(-0.182\) |
GB | 0.011 | 0.067 | 103.727 | 0.827 | 0.702 | |
KNN | 0.012 | 0.042 | 65.334 | 0.810 | 0.882 | |
LR | 0.017 | 0.100 | 154.280 | 0.746 | 0.340 | |
RF | 0.008 | 0.035 | 53.294 | 0.872 | 0.921 | |
L2 | 0.011 | 0.044 | 67.856 | 0.819 | 0.872 | |
XGB | 0.015 | 0.058 | 89.770 | 0.746 | 0.777 | |
LSTM | 0.017 ± 0.003 | 0.033 ± 0.003 | 50.673 ± 3.878 | 0.741 ± 0.038 | 0.929 ± 0.011 | |
LSTM_IPIP | 0.014 ± 0.003 | 0.038 ± 0.005 | 59.045 ± 7.008 | 0.784 ± 0.034 | 0.902 ± 0.024 | |
12H | DT | 0.074 | 0.383 | 589.773 | 0.165 | − 8.633 |
GB | 0.062 | 0.187 | 287.690 | 0.216 | − 1.292 | |
KNN | 0.057 | 0.326 | 502.114 | 0.335 | − 5.982 | |
LR | 0.058 | 0.149 | 230.323 | 0.217 | − 0.469 | |
RF | 0.068 | 0.229 | 353.007 | 0.198 | − 2.451 | |
L2 | 0.058 | 0.142 | 219.594 | 0.221 | − 0.336 | |
XGB | 0.078 | 0.214 | 330.679 | 0.197 | − 2.028 | |
LSTM | 0.028 ± 0.002 | 0.108 ± 0.005 | 165.773 ± 8.374 | 0.545 ± 0.012 | 0.236 ± 0.074 | |
LSTM_IPIP | 0.028 ± 0.003 | 0.104 ± 0.005 | 159.818 ± 7.251 | 0.548 ± 0.015 | 0.290 ± 0.064 | |
24H | DT | 1.956 | 4.258 | 6565.067 | 0.008 | − 1191.702 |
GB | 1.148 | 2.287 | 3526.053 | 0.014 | − 343.058 | |
KNN | 0.319 | 0.942 | 1452.420 | 0.056 | − 57.376 | |
LR | NC | NC | NC | NC | NC | |
RF | 1.212 | 2.499 | 3852.896 | 0.011 | − 409.798 | |
L2 | 0.039 | 0.124 | 190.880 | 0.297 | − 0.008 | |
XGB | 0.458 | 0.468 | 720.896 | 0.046 | − 13.381 | |
LSTM | 0.029 ± 0.001 | 0.118 ± 0.001 | 180.525 ± 0.873 | 0.494 ± 0.007 | 0.093 ± 0.009 | |
LSTM_IPIP | 0.028 ± 0.001 | 0.116 ± 0.001 | 177.702 ± 1.736 | 0.484 ± 0.022 | 0.121 ± 0.017 |