Table 7 Performance evaluation of the proposed method with rank aggregation ensemble feature selection without SMOTE.
From: Enhancing blockchain transaction classification with ensemble learning approaches
Dataset | Methodology | Acy | Pre | Rec | F-1 | Spe | BAcy | AUC |
|---|---|---|---|---|---|---|---|---|
D1 | Hard Voting | 77.86 | 84.59 | 86.96 | 85.76 | 47.93 | 67.45 | 0.753 |
Soft Voting | 76.89 | 85.93 | 83.39 | 84.64 | 55.90 | 69.65 | 0.717 | |
Weighted Averaging | 75.88 | 85.36 | 82.78 | 84.05 | 53.12 | 67.95 | 0.726 | |
D2 | Hard Voting | 86.33 | 85.46 | 88.73 | 87.06 | 83.76 | 86.24 | 0.781 |
Soft Voting | 84.50 | 83.18 | 87.39 | 85.23 | 81.48 | 84.43 | 0.795 | |
Weighted Averaging | 81.30 | 83.98 | 83.54 | 83.76 | 78.25 | 80.89 | 0.776 | |
D3 | Hard Voting | 93.95 | 95.89 | 96.44 | 96.16 | 84.85 | 90.64 | 0.912 |
Soft Voting | 93.05 | 94.93 | 96.42 | 95.67 | 79.90 | 88.16 | 0.901 | |
Weighted Averaging | 92.90 | 94.54 | 96.38 | 95.45 | 81.06 | 88.72 | 0.897 |
