Research Article
On Internet Traffic Classification: A Two-Phased Machine Learning Approach
Table 6
Feature sets versus classifier accuracy.
| Feature set | Pruning = false | Pruning = true | No boost | Boost 10 | Boost 100 | No boost | Boost 10 | Boost 100 |
| Set 1 | 39.58 | 40.01 | 41.34 | 39.44 | 40.48 | 41.97 | Set 2 | 24.29 | 24.29 | 24.29 | 24.29 | 24.29 | 24.29 | Set 3 | 82.29 | 83.24 | 84.29 | 82.20 | 84.97 | 83.95 | Set 4 | 73.18 | 75.51 | 75.70 | 73.18 | 72.62 | 75.03 | Sets 1 + 3 | 91.37 | 94.39 | 95.98 | 92.37 | 94.52 | 96.67 | Sets 1 + 4 | 84.48 | 87.47 | 86.47 | 84.48 | 86.42 | 86.79 | Sets 2 + 3 | 84.90 | 86.91 | 85.71 | 84.90 | 85.00 | 85.61 | Sets 2 + 4 | 74.37 | 77.07 | 77.21 | 74.37 | 76.83 | 77.42 |
|
|