Table 2 Classification performance using independent measures.
Study | Measures used | Classifier | Parameters tuned | Accuracy | Sensitivity | Specificity | F1 | AUC |
|---|---|---|---|---|---|---|---|---|
1 | Subcortical volumes | Support Vector Classifier | Regularization parameter C, Max iterations | 72% | 79% | 57% | 0.79 | 0.68 |
2 | Cortical volumes | Nu-Support Vector Classifier | Regularization parameter C, Max iterations, Kernel | 73% | 81% | 57% | 0.73 | 0.69 |
3 | Cortical surface areas | Support Vector Classifier | Regularization parameter C, Max iterations | 73% | 77% | 65% | 0.74 | 0.72 |
4 | Cortical thickness | Support Vector Classifier | Regularization parameter C, Max iterations | 75% | 81% | 61% | 0.75 | 0.71 |
5 | Cortical mean curvature | Logistic Regression | Regularization parameter C, Max iterations, penalty, solver | 70% | 73% | 65% | 0.71 | 0.69 |
