Table 7.
Group Classification | Dataset | GLM | SSM/PCA1 | SSM/PCA2 | SVM-ISDA | SVM-SMO |
---|---|---|---|---|---|---|
$AD vs. NL | ADNI | 0.675 | 0.626 | 0.703 | *0.803 | 0.797 |
stable MCI vs. prodromal AD | ADNI | 0.472 | 0.406 | 0.493 | *0.514 | 0.477 |
Winnipeg | *0.590 | 0.550 | 0.495 | *0.590 | 0.534 | |
Busan (SPECT) | 0.000 | 0.000 | 0.000 | 0.525 | *0.563 |
*The best performance.
$Since this is the training set of the classification machines, the AD designation was based on 10-fold cross-validation.