Table 2 Classifier rankings based on average ranks across testing datasets and noise intensities (0% to 10%) highlighting NoiseCut as the best-performing method
Noise Intensity | Model | The Average Rank Across All Experiments |
|---|---|---|
0% | NoiseCut | 1.540 |
DNN | 2.670 | |
XGBoost | 2.930 | |
SVM | 3.466 | |
RF | 4.393 | |
2.5% | NoiseCut | 1.490 |
DNN | 2.340 | |
XGBoost | 2.753 | |
SVM | 3.416 | |
RF | 5.000 | |
5% | NoiseCut | 1.506 |
DNN | 2.316 | |
XGBoost | 2.730 | |
SVM | 3.446 | |
RF | 5.000 | |
7.5% | NoiseCut | 1.543 |
DNN | 2.473 | |
XGBoost | 2.730 | |
SVM | 3.253 | |
RF | 5.000 | |
10% | NoiseCut | 1.553 |
DNN | 2.503 | |
XGBoost | 2.676 | |
SVM | 3.266 | |
RF | 5.000 |