Table 2 Hyperparameter optimization results of well H21.
From: Steering drilling wellbore trajectory prediction based on the NOA-LSTM-FCNN method
Method | Number of hidden layers | Number of convolution layers | Maximum training times | Minimum training batch | learning rate |
|---|---|---|---|---|---|
BP | 32 | \ | 1000 | 12 | 0.0257 |
CNN | 48 | 133 | 800 | 16 | 0.0156 |
LSTM | 35 | \ | 1000 | 35 | 0.05613 |
GRU | 76 | \ | 742 | 29 | 0.02437 |
LSTM-FCNN | 64 | 251 | 893 | 41 | 0.08421 |
