Table 5 Hyperparameters of the PCPFN model after algorithmic optimization.
From: An efficient parallel runoff forecasting model for capturing global and local feature information
Range | Dongjiang | Quinebaug | Housatonic | |
|---|---|---|---|---|
BiGRU_hidden_size | [2,10] | 297 | 99 | 261 |
BiGRU_layers_num | [2,10] | 1 | 2 | 2 |
Channel_embedding | [5,300] | 16 | 112 | 16 |
Nhead | [2,10] | 8 | 8 | 8 |
Feedforward_dim | [5,300] | 20 | 213 | 152 |
Encoder_layers_num | [1,5] | 1 | 1 | 2 |
Encoder_dropout_rate | [0,0.3] | 0.28 | 0.27 | 0.18 |
Linear_dropout_rate | [0,0.3] | 0.23 | 0.25 | 0.2 |
Alpha | [0.3*L,3*L] | 1660.52 | 2332.21 | 2111.12 |
K | [2,10] | 10 | 5 | 8 |
Init | [0,3] | 1 | 0 | 1 |
