Table 5. Model error decimal table.
From: Neutron transport calculation for the BEAVRS core based on the LSTM neural network
Serial number | Time step | Number of units | Regularization factor | Optimizer | Average error/pcm |
|---|---|---|---|---|---|
1 | 3 | 16 | 0.003 | Adadelta | 0.5266 |
2 | 1 | 16 | 0.001 | Adadelta | 0.9370 |
3 | 2 | 4 | 0.002 | Adadelta | 0.9514 |
4 | 1 | 4 | 0.001 | RMSProp | 0.9742 |
5 | 1 | 8 | 0.001 | Adadelta | 1.1812 |
6 | 2 | 8 | 0.004 | Adadelta | 1.2292 |
7 | 1 | 16 | 0.001 | RMSProp | 1.5629 |
8 | 2 | 8 | 0.005 | Adadelta | 1.7697 |
9 | 2 | 8 | 0.001 | Adadelta | 1.7911 |
10 | 1 | 8 | 0.002 | RMSProp | 1.9308 |
