Research Article
Standard Precipitation Index Drought Forecasting Using Neural Networks, Wavelet Neural Networks, and Support Vector Regression
Table 3
(a) Performance results for the Ejersalele station, Upper Awash Basin, (b) Performance results of Nazereth Station, Middle Awash Basin, (c) Performance results of Dubti Station, Lower Awash Basin.
(a) |
| Model-Lead time | SPI 3 | SPI 12 | | RMSE | MAE | | RMSE | MAE |
| ANN-L1 | 0.7694 | 0.1574 | 0.1433 | 0.9451 | 0.0610 | 0.0603 | ANN-L6 | 0.6232 | 0.1744 | 0.1567 | 0.8614 | 0.1011 | 0.0885 | WN-L1 | 0.8829 | 0.0700 | 0.0352 | 0.9534 | 0.0600 | 0.0536 | WN-L6 | 0.6433 | 0.1070 | 0.0356 | 0.8731 | 0.0790 | 0.0662 | SVR-L1 | 0.7219 | 0.1046 | 0.0915 | 0.7611 | 0.1312 | 0.1129 | SVR-L6 | 0.6647 | 0.1118 | 0.1042 | 0.6941 | 0.1341 | 0.1247 |
|
|
(b) |
| Model-Lead time | SPI 3 | SPI 12 | | RMSE | MAE | | RMSE | MAE |
| ANN-L1 | 0.7319 | 0.1170 | 0.1016 | 0.9158 | 0.1003 | 0.0911 | ANN-L6 | 0.6546 | 0.1240 | 0.1142 | 0.7542 | 0.1104 | 0.0919 | WN-L1 | 0.9483 | 0.0510 | 0.0441 | 0.9167 | 0.0753 | 0.0629 | WN-L6 | 0.8641 | 0.0727 | 0.0512 | 0.8012 | 0.1072 | 0.0802 | SVR-L1 | 0.7114 | 0.1216 | 0.1114 | 0.7713 | 0.1147 | 0.1130 | SVR-L6 | 0.6540 | 0.1320 | 0.1217 | 0.7326 | 0.1244 | 0.1215 |
|
|
(c) |
| Model-Lead time | SPI 3 | SPI 12 | | RMSE | MAE | | RMSE | MAE |
| ANN-L1 | 0.7368 | 0.1175 | 0.1095 | 0.9188 | 0.0710 | 0.0648 | ANN-L6 | 0.6806 | 0.1302 | 0.1147 | 0.7135 | 0.0938 | 0.0836 | WN-L1 | 0.9018 | 0.0652 | 0.0581 | 0.9473 | 0.0648 | 0.0560 | WN-L6 | 0.8119 | 0.0706 | 0.0642 | 0.8641 | 0.0846 | 0.0747 | SVR-L1 | 0.6990 | 0.1146 | 0.1022 | 0.7041 | 0.1102 | 0.1009 | SVR-L6 | 0.6331 | 0.1309 | 0.1242 | 0.6705 | 0.1107 | 0.1025 |
|
|