Research Article
Feature Selection Based on Cross-Correlation for the Intrusion Detection System
Table 2
Improvement percentage of the proposed method (CCFS) in comparison with the CFA method for the DT classifier.
| KDD Cup 99 | Number of features | CCFS improvement versus CFA (%) | Accuracy | Precision | Recall | F1-score |
| 41 | 1.42 | 3.40 | 2.19 | 2.79 | 35 | 2.05 | 2.61 | 2.61 | 2.61 | 30 | 0.72 | 0.82 | 0.16 | 0.49 | 25 | 1.83 | 3.70 | 3.35 | 3.52 | 20 | 0.52 | 0.99 | 2.16 | 1.57 | 15 | 0.48 | 0.13 | 0.51 | 0.32 | 10 | 0.20 | 1.62 | 1.62 | 1.62 | 5 | 1.06 | 2.50 | 1.86 | 2.18 |
|
|