Multi-Model Ensemble Learning for Encrypted Traffic Detection
Abstract
In encrypted traffic detection, single classifiers suffer from weak generalization and limited feature representation. To solve this, a multi-model ensemble method is proposed. First, Random Forest, XGBoost, and Decision Tree form a heterogeneous pool. Second, weights are finely adjusted via a sliding window mechanism. Finally, the optimal weight combination is identified by dynamic search. Experimental results show that the proposed method effectively enhances detection accuracy and robustness on imbalanced datasets.
Fields of Study & Subjects
Internet Traffic Analysis and Secure E-voting
Network Security and Intrusion Detection
Imbalanced Data Classification Techniques
Publication details
- Journal / source
- International Conference on Consumer Electronics and Computer Engineering (ICCECE)
- Pages
- 113–116
- Publication date
- 23 January 2026
- Language
- EN