Anomaly based intrusion detection using meta ensemble classifier
Proceedings of the Fifth International Conference on Security of Information and Networks
| Hi-index | 0.00 |
Recently, DoS (Denial of Service) detection has become more and more important in web security. In this paper, we argue that DoS attack can be taken as continuous data streams, and thus can be detected by using stream data mining methods. More specifically, we propose a new Weighted Ensemble learning model to detect the DoS attacks. The Weighted Ensemble model first trains base classifiers using different data classification algorithms (i.e., decision tree, SVMs, and Naive Bayes) on multiple successive data chunks, and then weights each base classifier according to its prediction accuracy on the up-to-date data. Experimental results on the benchmark KDDCUP’99 dataset demonstrate that our new Weighted Ensemble model is able to successfully detect DoS attacks.