Research Article
Automatic Sleep Stage Classification Based on Convolutional Neural Network and Fine-Grained Segments
Table 3
Evaluation results of the original CNN with 5 channels.
| Networks | Stages | Precision | Sensitivity | Specificity | Accuracy | Best | Worst | Best | Worst | Best | Worst | Best | Worst |
| CNN | W | 0.9510 | 0.7501 | 0.9616 | 0.5 | 0.9964 | 0.7997 | 0.9009 | 0.8078 | N1 | 0.8168 | 0.4284 | 0.8219 | 0.71 | 0.9807 | 0.7518 | 0.7928 | 0.6917 | N2 | 0.9110 | 0.6098 | 0.9033 | 0.3182 | 1 | 0.6974 | 0.9034 | 0.7560 | N3 | 0.9636 | 0.6028 | 0.9635 | 0.5657 | 1 | 0.8504 | 0.9724 | 0.7588 | REM | 0.9229 | 0.593 | 0.8351 | 0.5258 | 0.9972 | 0.8067 | 0.9329 | 0.8067 | Average | 0.8937 | 0.7669 | 0.9192 | 0.8325 |
|
|