Research Article
A Dense Long Short-Term Memory Model for Enhancing the Imagery-Based Brain-Computer Interface
Table 2
Influence of the unit size of network parameters in each layer.
| Parameter unit | Accuracy (%) |
| 128-256 (LSTM-DENSE) | 82.32 | 128-128 (LSTM-DENSE) | 87.66 | 128-64 (LSTM-DENSE) | 88.93 | 128-32 (LSTM-DENSE) | 89.52 | 128-256-256 (LSTM-DENSE-DENSE) | 85.43 | 128-256-128 (LSTM-DENSE-DENSE) | 86.42 | 128-256-64 (LSTM-DENSE-DENSE) | 88.59 | 128-256-32 (LSTM-DENSE-DENSE) | 89.32 | 128-128-64 (LSTM-DENSE-DENSE) | 89.45 | 128-128-32 (LSTM-DENSE-DENSE) | 90.22 | 128-64-32 (LSTM-DENSE-DENSE) (MSE) | 90.68 | 128-64-32 (LSTM-DENSE-DENSE) (Softmax) | 91.56 |
|
|