| | Classification | Type of network | Type of layer | Output sequence | Number of parameters | Total number of parameters |
| | Pure RNN | Pure GRU | GRU | (None, 41, 64) | 12864 | 15,009 | | Dense | (None, 32) | 1056 | | Dense | (None, 33) | 1089 | | Pure LSTM | LSTM | (None, 64) | 16896 | 20,065 | | Dense | (None, 32) | 2080 | | Dense | (None, 33) | 1089 |
| | Multi-input hybrid model | GRU-BP | Input layer | [(None, 32, 1)] | 0 | 25,089 | | Input layer | [(None, 9)] | 0 | | GRU | (None, 32, 64) | 12864 | | Dense | (None, 16) | 160 | | GRU | (None, 32) | 9408 | | Concatenate | (None, 48) | 0 | | Dense | (None, 32) | 1568 | | Dense | (None, 33) | 1089 | | LSTM-BP | Input layer | [(None, 32, 1)] | 0 | 32,129 | | Input layer | [(None, 9)] | 0 | | LSTM | (None, 32, 64) | 16896 | | Dense | (None, 16) | 160 | | LSTM | (None, 32) | 12416 | | Concatenate | (None, 48) | 0 | | Dense | (None, 32) | 1568 | | Dense | (None, 33) | 1089 | | Stacking models | LSTM-Bi | Bidirectional (LSTM (64)) | (None, 41, 128) | 33792 | 78,177 | | Bidirectional (LSTM (32)) | (None, 64) | 41216 | | Dense | (None, 32) | 2080 | | Dense | (None, 33) | 1089 | | LTSM-Stack | LSTM | (None, 41, 256) | 264192 | 525,281 | | LSTM | (None, 41, 128) | 197120 | | LSTM | (None, 41, 64) | 49408 | | LSTM | (None, 32) | 12416 | | Dense | (None, 32) | 1056 | | Dense | (None, 33) | 1089 |
| | Space-time model | ConvLSTM | ConvLSTM2D | (None, 41, 1, 1, 41) | 62156 | 117,233 | | BN | (None, 41, 1, 1, 41) | 164 | | Flatten | (None, 1681) | 0 | | Dense | (None, 32) | 53824 | | Dense | (None, 33) | 1089 |
|
|
