Table 2 Accuracy and precision for the Ecoli dataset.
From: Algorithmic and mathematical modeling for synthetically controlled overlapping
Approach | Accuracy | Precision | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
 Method | 0% | 10% | 20% | 30% | 40% | 50% | 0% | 10% | 20% | 30% | 40% | 50% |
Majority class overlapping scheme (MOS) | Â | |||||||||||
 KNN | 86.76 | 85.71 | 72.60 | 67.11 | 62.03 | 60.98 | 86.46 | 86.54 | 73.57 | 68.93 | 63.26 | 59.12 |
 SVM | 88.24 | 80.00 | 79.45 | 77.63 | 67.79 | 65.73 | 87.05 | 81.09 | 80.43 | 80.41 | 75.46 | 70.81 |
 RF | 91.18 | 81.43 | 79.45 | 73.68 | 63.29 | 68.29 | 89.84 | 81.87 | 79.15 | 74.48 | 69.04 | 66.36 |
All class overlapping scheme (AOS) | ||||||||||||
 KNN | 86.76 | 75.68 | 65.43 | 52.59 | 48.42 | 44.55 | 86.46 | 80.29 | 64.78 | 52.63 | 48.19 | 45.97 |
 SVM | 88.24 | 85.14 | 72.84 | 66.23 | 63.21 | 59.39 | 87.05 | 85.98 | 76.17 | 68.08 | 64.16 | 55.16 |
 RF | 91.18 | 81.08 | 69.14 | 62.50 | 52.63 | 46.53 | 89.94 | 83.22 | 72.26 | 69.90 | 58.99 | 51.85 |
Random class overlapping scheme (ROS) | ||||||||||||
 KNN | 86.76 | 83.82 | 83.82 | 87.14 | 85.29 | 65.85 | 86.46 | 84.67 | 85.16 | 87.53 | 85.55 | 63.74 |
 SVM | 88.24 | 85.29 | 82.35 | 88.57 | 85.29 | 70.73 | 87.05 | 85.15 | 86.01 | 90.65 | 87.33 | 75.81 |
 RF | 86.76 | 80.88 | 83.82 | 84.29 | 86.76 | 59.51 | 85.54 | 82.07 | 86.88 | 89.49 | 87.66 | 70.33 |
SMOTE and all class overlapping scheme (AOS-SMOTE) | ||||||||||||
 KNN | 96.25 | 93.53 | 94.47 | 92.02 | 87.97 | 88.93 | 96.36 | 93.64 | 94.84 | 92.21 | 87.97 | 89.37 |
 SVM | 94.32 | 91.38 | 92.77 | 94.12 | 87.55 | 89.75 | 94.37 | 91.47 | 93.05 | 94.18 | 87.70 | 91.34 |
 RF | 97.38 | 95.26 | 94.47 | 93.28 | 90.46 | 89.75 | 97.53 | 95.23 | 95.09 | 93.62 | 90.49 | 89.75 |
