Table 3.
Optimal model using multiple channels—confusion matrix.
| Predicted Classes | True Positive Rate | False Negative Rate | ||||
|---|---|---|---|---|---|---|
| Resting State (CHT Does Not Exist) L = (1;0;0) |
Picking (CHT = 0) L = (0;1;0) |
Coordination (CHT > 0) L = (0;0;1) |
||||
| True Classes | Resting state (CHT does not exist) L = (1;0;0) |
589 | 11 | 0 | 98.17% | 1.83% |
| Picking (CHT = 0) L = (0;1;0) |
3 | 89 | 28 | 74.16% | 25.84% | |
| Coordination (CHT > 0) P = (0;0;1) |
0 | 28 | 92 | 76.67% | 23.33% | |
| Positive Predicted Value | 99.49% | 69.53% | 72.44 % | Total Prediction Accuracy (770/840) 91.66% |
||
| False Discovery Rate | 1.51% | 30.47% | 27.56% | |||