Research Article
Design of an Evolutionary Approach for Intrusion Detection
Table 6
Overview of classification results of KDD cup 1999 subsets using NB as a base classifier.
| Dataset | Technique | Avg. DR | Avg. FPR | CID | Normal | Probe | DoS | U2R | R2L |
| KDD 1 | NB | 0.619 | 0.208 | 0.129 | 0.698 | 0.960 | 0.387 | 0.140 | 0.140 | Bagged-NB | 0.651 | 0.219 | 0.140 | 0.746 | 0.960 | 0.387 | 0.140 | 0.140 | Boosted-NB | 0.619 | 0.208 | 0.129 | 0.698 | 0.960 | 0.387 | 0.140 | 0.140 | AMGA2-NB | 0.736 | 0.260 | 0.165 | 0.872 | 0.893 | 0.413 | 0.140 | 0.220 |
| KDD 2 | NB | 0.549 | 0.085 | 0.157 | 0.691 | 0.939 | 0.449 | 0.180 | 0.281 | Bagged-NB | 0.549 | 0.085 | 0.157 | 0.691 | 0.939 | 0.449 | 0.180 | 0.281 | Boosted-NB | 0.548 | 0.085 | 0.157 | 0.691 | 0.939 | 0.449 | 0.170 | 0.280 | AMGA2-NB | 0.616 | 0.091 | 0.194 | 0.820 | 0.945 | 0.450 | 0.200 | 0.461 |
| ITFS KDD 41 features | NB | 0.446 | 0.120 | 0.074 | 0.945 | 0.972 | 0.353 | 0.254 | 0.223 | Bagged-NB | 0.442 | 0.122 | 0.071 | 0.944 | 0.957 | 0.351 | 0.241 | 0.221 | Boosted-NB | 0.446 | 0.120 | 0.074 | 0.945 | 0.972 | 0.353 | 0.254 | 0.223 | AMGA2-NB | 0.604 | 0.060 | 0.197 | 0.855 | 0.997 | 0.287 | 0.145 | 0.814 |
| ITFS KDD 10 features | NB | 0.566 | 0.233 | 0.067 | 0.775 | 0.718 | 0.657 | 0.171 | 0.326 | Bagged-NB | 0.540 | 0.237 | 0.056 | 0.775 | 0.717 | 0.599 | 0.158 | 0.326 | Boosted-NB | 0.566 | 0.233 | 0.067 | 0.775 | 0.718 | 0.657 | 0.171 | 0.326 | AMGA2-NB | 0.703 | 0.105 | 0.226 | 0.807 | 0.896 | 0.615 | 0.118 | 0.731 |
|
|