Table 3 Classification accuracy and details of the proposed method and deep neural network approaches which used raw EEG signals11,12,17,21,23,25,26 for the SEED dataset.
Subject-dependent/independent | Methods | Used signals | Accuracy (%) |
|---|---|---|---|
Subject-dependent | Jin et al.26 (2020) | Raw EEG | 98.51 |
Zhong et al.25 (2020) | Raw EEG | 92.58 | |
Khare et al.23 (2020) | Raw EEG | 91.10 | |
Wang et al.21 (2020) | Raw EEG | 89.62 | |
Song et al.17 (2018) | Raw EEG | 84.43 | |
Chen et al.12 (2020) | Raw EEG | 81.47 | |
Padilla et al.11 (2016) | Raw EEG | 62.95 | |
Proposed method | EEG source signal | 99.25 | |
Subject-independent | Jin et al.26 (2020) | Raw EEG | 97.77 |
Zhong et al.25 (2020) | Raw EEG | 88.88 | |
Khare et al.23 (2020) | Raw EEG | 81.48 | |
Wang et al.21 (2020) | Raw EEG | 74.07 | |
Song et al.17 (2018) | Raw EEG | 72.59 | |
Chen et al.12 (2020) | 70.37 | ||
Padilla et al.11 (2016) | 59.25 | ||
Proposed method | EEG source signal | 98.51 |
