Table 3.
Accuracy obtained with/without the use of the feature selection method, for each sub-dataset.
| Sub-Dataset | Features Selection | Supervised | Unsupervised | ||||||
|---|---|---|---|---|---|---|---|---|---|
| Performances | K-NN | CART | RF | NB | SVM | K-Means | GMM | ||
| Yogev et al. | With | Accuracy | 86.05% | 83.72% | 86.05% | 74.42% | 86.05% | 63.72% | 64.77% |
| Without | Accuracy | 82.56% | 80.23% | 84.88% | 72.09% | 82.56% | 53.72% | 61.39% | |
| Hausdorff et al. | With | Accuracy | 90.91% | 84.30% | 87.60% | 77.69% | 90.08% | 55.12% | 65.12% |
| Without | Accuracy | 88.43% | 82.64% | 85.95% | 72.73% | 87.60% | 51.07% | 58.93% | |
| Frenkel-Toledo et al. | With | Accuracy | 81.25% | 79.69% | 82.81% | 79.69% | 82.81% | 57.19% | 65.31% |
| Without | Accuracy | 79.69% | 76.56% | 78.12% | 75% | 81.25% | 53.75% | 57.34% | |