Table 3 Comparison with multiple CSP-based feature extraction methods
From: A dynamic directed transfer function for brain functional network-based feature extraction
Reference number | Methods | Subjects | Average accuracies (%) | ||
|---|---|---|---|---|---|
‘k3b’ | ‘k6b’ | ‘l1b’ | |||
[7] | AC-CSP | 97.80 | 63.30 | 94.20 | 85.10 |
[7] | AC-RCSP | 97.80 | 72.50 | 95.00 | 88.43 |
[7] | CCS-CSP | 98.90 | 79.20 | 95.80 | 91.30 |
[7] | CCS-RCSP | 98.90 | 80.00 | 96.70 | 91.87 |
[8] | CSSSP | 95.50 | 55.10 | 95.00 | 81.87 |
[8] | BCSP | 78.80 | 63.70 | 76.60 | 73.03 |
[8] | ACSP | 76.60 | 56.80 | 51.60 | 61.67 |
[8] | Pcv | 100.00 | 68.90 | 96.60 | 88.50 |
[8] | Pfix | 100.00 | 67.20 | 98.30 | 88.50 |
[9] | CSP | 67.50 | 70.00 | 53.30 | 63.60 |
[9] | CCSP | 95.00 | 90.00 | 83.30 | 89.43 |
This paper | DDTF | 100.00 | 82.25 | 99.75 | 94.00 |