Table 2.
Classifier performance comparison over all designed features.
| Method | Sensitivity | Specificity | AUC | F1 | MCC | |
|---|---|---|---|---|---|---|
| 1 | Support Vector Machines | 0.926 | 0.945 | 0.935 | 0.901 | 0.859 |
| 2 | Random Forest | 0.870 | 0.957 | 0.914 | 0.883 | 0.836 |
| 3 | Linear Discriminant Analysis | 0.935 | 0.919 | 0.927 | 0.881 | 0.830 |
| 4 | Logistic Regression | 0.875 | 0.941 | 0.974 | 0.867 | 0.816 |
| 5 | Decision Tree | 0.861 | 0.943 | 0.902 | 0.863 | 0.808 |
| 6 | Naive Bayes | 0.875 | 0.894 | 0.884 | 0.824 | 0.746 |
Each classifier performance was evaluated using five metrics: Sensitivity, Specificity, Area Under Curve (AUC), F1-Score and MCC. Results are sorted by decreasing value of F1 and MCC.