TABLE 5.
Performance of the best models developed on the HSE vs INA partitioning scheme on the TeS. For each model, the following statistics are reported: Balanced Accuracy (BA), Sensitivity (SEN), Specificity (SPE), Matthews Correlation Coefficient (MCC), number of True Negatives (TN), False Positives (FP), True Positives (TP) and False Negatives (FNs).
| Classifier | BRF | KNN | SVM | RF |
|---|---|---|---|---|
| #Descriptors | 160 | 20 | 50 | 20 |
| BA | 0.85 | 0.86 | 0.79 | 0.78 |
| SEN | 0.84 | 0.87 | 0.68 | 0.65 |
| SPE | 0.87 | 0.85 | 0.91 | 0.92 |
| MCC | 0.61 | 0.60 | 0.56 | 0.56 |
| TN | 130 | 127 | 136 | 138 |
| FP | 20 | 23 | 14 | 12 |
| TP | 26 | 27 | 21 | 20 |
| FN | 5 | 4 | 10 | 11 |
| # | 181 | 181 | 181 | 181 |