Table 9 Comparison with other methods
From: A novel two-phase cycle algorithm for effective cyber intrusion detection in edge computing
Method | Normal | DoS | Probe | U2R | R2L | Accuracy | DR | FAR |
|---|---|---|---|---|---|---|---|---|
Multi-level SVM-ELM [45] | 98.13 | 99.54 | 87.22 | 21.93 | 31.39 | 95.75 | 95.17 | 1.87 |
NFC[36] | 98.2 | 99.5 | 84.1 | 14.1 | 31.5 | N/A | 95.2 | 1.9 |
Genetical algorithm [45] | 69.5 | 99.4 | 71.1 | 18.9 | 5.4 | 90 | 94.95 | 30.46 |
SVM+BIRCH clustering [45] | 99.3 | 99.5 | 97.5 | 19.7 | 28.8 | 95.7 | N/A | 0.7 |
MOGFIDS [45] | 98.36 | 97.2 | 88.6 | 15.79 | 11.01 | 93.2 | 91.96 | 1.6 |
Association rules [45] | 99.47 | 96.6 | 74.8 | 3.8 | 1.21 | 92.4 | N/A | 0.53 |
Multiclass SVM [45] | 99.6 | 96.8 | 75 | 5.3 | 4.2 | 92.46 | 90.74 | 0.43 |
Winning the KDD99 [45] | 99.5 | 97.1 | 83.3 | 13.2 | 8.4 | 93.3 | 91.81 | 0.55 |
t-SNE SVM [46] | 99.85 | 99.62 | 98.89 | 86.8 | 96.51 | N/A | N/A | N/A |
Hybrid classifier [47] | 96.8 | 98.66 | 93.4 | 46.97 | 71.43 | 96.77 | 96.77 | 3 |
DT-SVM [48] | 99.7 | 99.92 | 98.57 | 48 | 37.8 | N/A | N/A | N/A |
GA FPM [49] | N/A | 98.85 | 98.05 | 78.7 | 98.7 | N/A | N/A | N/A |
Proposed approach (combined) | 99.28 | 98.36 | 96.71 | 100 | 86 | 98.81 | 98.23 | 0.8 |