Table 2 Performance comparison of algorithms in Weka workbench on the KDD99 dataset [4]
From: Foundations and applications of artificial Intelligence for zero-day and multi-step attack detection
Algorithm | Percentage of successful prediction (%) | Training time (seconds) |
|---|---|---|
K-Means | 78.7 | 70.7 |
NEA | 92.22 | 10.63 |
FCC | 89.2 | 56.2 |
ID3 | 72.22 | 120 |
J48 | 92.06 | 15.85 |
PART | 45.67 | 169 |
NBTree | 92.28 | 25.88 |
SVM | 81.38 | 222.28 |
Fuzzy logic | 91.8 | 873.9 |
Naive Bayes | 78.32 | 5.57 |
BayesNet | 90.62 | 6.28 |
Decision Table | 91.66 | 66.24 |
Random Forest Classifier | 92.81 | 491 |
Jrip | 92.30 | 207.47 |
OneR | 89.31 | 3.75 |
MLP | 92.03 | 350.15 |
SOM | 91.65 | 192.16 |
GAU | 69.9 | 177.4 |
MARS | 96.5 | 67.9 |
Apriori | 87.5 | 18 |