Research Article
A Comparative Study of VMD-Based Hybrid Forecasting Model for Nonstationary Daily Streamflow Time Series
Table 3
Performance of the GBRT model during the training period at the Weijiabao station.
| | Sequence | GBRT parameters | Training | | LR | MD | MF | MSS | MSL | RMSE (m3/s) | NSE |
| | IMF1 | 0.17 | 9 | 10 | 170 | 63 | 3.311 | 0.99831 | | IMF2 | 0.37 | 15 | 4 | 200 | 1 | 0.452 | 0.99992 | | IMF3 | 0.27 | 15 | 7 | 200 | 1 | 0.624 | 0.99972 | | IMF4 | 0.15 | 6 | 3 | 130 | 37 | 5.748 | 0.96849 | | IMF5 | 0.25 | 15 | 1 | 34 | 5 | 1.070 | 0.99788 | | IMF6 | 1 | 15 | 7 | 200 | 1 | 0.137 | 0.99996 | | IMF7 | 0.20 | 15 | 7 | 2 | 1 | 0.042 | 0.99999 | | IMF8 | 1 | 1 | 7 | 200 | 1 | 2.049 | 0.97926 | | IMF9 | 0.13 | 15 | 6 | 2 | 1 | 0.040 | 0.99999 | | IMF10 | 0.15 | 6 | 4 | 130 | 37 | 2.912 | 0.94075 | | IMF11 | 1 | 15 | 3 | 200 | 1 | 0.101 | 0.99990 |
|
|
