Research Article
An Intrusion Detection Method Based on Decision Tree-Recursive Feature Elimination in Ensemble Learning
| Input: Training sample set | (1) | Initialize original feature set and feature ordering set []. | (2) | fordo | (3) | The Decision Tree classifier is trained, and the feature selection of single variable by F-test (ANOVA) is obtained. | (4) | Calculate ranking criterion score | (5) | Find the feature with the lowest ranking score | (6) | Update feature set R = [p, R) | (7) | Remove other features in S: S = S/p | (8) | until s = [ ] | (9) | end for | | output: Feature sort set R |
|