Table 4 Comparison between our experiment and other sleep stage classification methods
From: An automatic method using MFCC features for sleep stage classification
Article | Dataset | Method | Channel | Subjects | ACC (%) | \(\kappa\) | F1-score | ||||
|---|---|---|---|---|---|---|---|---|---|---|---|
W | S1 | S2 | S3 | REM | |||||||
Phan et al. [15] | Sleep-EDF | Multitask 1-max CNN | Fpz-Cz | 20 | 81.9 | 0.74 | – | – | – | – | – |
Qu et al. [20] | Sleep-EDF | CNN | Fpz-Cz | 20 | 84.3 | 0.78 | 90.2 | 48.3 | 87.8 | 85.6 | 83.0 |
Supratak et al. [38] | Sleep-EDF | DeepSleep- Net | Fpz-Cz | 20 | 82.0 | 0.76 | 84.7 | 46.6 | 85.9 | 84.8 | 82.4 |
Sors et al. 39 | SHHS1 | CNN | C4-A1 | 5728 | 86.8 | 0.81 | 91.4 | 42.7 | 88.0 | 84.9 | 85.4 |
Seo et al. 40 | SHHS1 | IITNet | C4-A1 | 5728 | 83.6 | 0.77 | 88.7 | 21.3 | 86.1 | 84.9 | 78.1 |
Eldele et al. 41 | SHHS1 | AttnSleep | C4-A1 | 329 | 84.2 | 0.78 | 86.7 | 33.2 | 87.1 | 87.1 | 82.1 |
This study | SHHS1 | CNN+LSTM | C4-A1, C3-A2, EMG | 100 | 82.4 | 0.75 | 93.8 | 27.1 | 79.5 | 64.1 | 82.2 |