A Self-Learning Network Anomaly Detection System using Majority Voting |
Network traffic is constantly changing. However, many current Intrusion Detection Systems require somewhat static conditions in order to work properly. In this paper, we propose ongoing training and updating procedures and introduce a self-learning Anomaly Detection System based on majority voting that can adapt to network changes by steadily exchanging small parts of training data and evaluate the performance of different replacement strategies for this process.
Hock D, Kappes M