Table 8.
Performance evaluations for three-category classification of task 4.
| Task | C | γ | Indicator type | Accuracy in test set | Precision in test set | Recall in test set | AUC_ROC in test set |
|---|---|---|---|---|---|---|---|
| 4 | 50 | 0.5 | class0 | 0.808 | 0.828 | 0.750 | 0.903 |
| class1 | 0.772 | 0.830 | 0.926 | ||||
| class2 | 0.825 | 0.855 | 0.945 | ||||
| Micro∗ | 0.808 | 0.808 | 0.923 | ||||
| Macro∗∗ | 0.808 | 0.812 | 0.925 |
∗Micro evaluators are globally calculated by counting the total true positives, false negatives, and false positives. ∗∗Macro evaluators are directly obtained by unweighted averaging metrics from all classes [40].