Research Article
A Hybrid Swarm Intelligence Algorithm for Intrusion Detection Using Significant Features
Table 8
Performance comparison of classification algorithms on accuracy rate.
| Classification Algorithms | Average accuracy (%) | Feature selection method |
|
C4.5 [6] | 99.11 | All features | 98.69 | Genetic algorithm | 98.84 | Best-first | 99.41 | Correlation feature selection |
|
BayesNet [6] | 99.53 | All features | 99.52 | Genetic algorithm | 98.91 | Best-first | 98.92 | Correlation feature selection |
| ABC-SVM [7] | 92.768 | Binary ABC | PSO-SVM [7] | 83.88 | GA-SVM [7] | 80.73 |
|
KNN [8] | 98.24 | All features | 98.11 | Fast feature selection |
|
Bayes Classifier [8] | 76.09 | All features | 71.94 | Fast feature selection |
| ANN [9] | 81.57 | Feature reduction |
| SSO-RF [10, 11] | 92.7 | SSO |
| Hybrid SSO [12] | 97.67 | SSO |
| RSDT [13] | 97.88 | Rough set |
| ID3 [13] | 97.665 | All features | C4.5 [13] | 97.582 |
| FC-ANN [14] | 96.71 | All features |
| Proposed MABC-EPSO | 88.59 | All features | 99.32 | Single feature selection method | 99.82 | Random feature selection method |
|
|