Table 3.
Modality | Classifier | AUC | ACC | Sensitivity | Specificity | F1-score |
---|---|---|---|---|---|---|
sMRI | LSVM | 0.562 | 0.556 | 0.515 | 0.596 | 0.532 |
RBFSVM | 0.564 | 0.544 | 0.473 | 0.616 | 0.501 | |
RF | 0.537 | 0.533 | 0.497 | 0.570 | 0.511 | |
dMRI | LSVM | 0.600 | 0.577 | 0.446 | 0.710 | 0.503 |
RBFSVM | 0.577 | 0.567 | 0.483 | 0.650 | 0.518 | |
RF | 0.568 | 0.552 | 0.528 | 0.575 | 0.534 | |
rsfMRI | LSVM | 0.622 | 0.603 | 0.531 | 0.675 | 0.567 |
RBFSVM | 0.655 | 0.616 | 0.602 | 0.628 | 0.605 | |
RF | 0.653 | 0.617 | 0.595 | 0.639 | 0.603 | |
Multimodal | LSVM | 0.664 | 0.629 | 0.565 | 0.696 | 0.598 |
RBFSVM | 0.668 | 0.611 | 0.599 | 0.623 | 0.602 | |
RF | 0.636 | 0.596 | 0.590 | 0.602 | 0.590 | |
MKL | 0.698 | 0.643 | 0.609 | 0.676 | 0.626 |
The bold values are the best AUC in all unimodal settings and the globally best metrics.