期刊名称：International Journal of Computer Science Issues
印刷版ISSN：1694-0784
电子版ISSN：1694-0814
出版年度：2018
卷号：15
期号：5
出版社：IJCSI Press
摘要：Distributed Denial of Service (DDoS) attacks considered the most critical attack for cyber security and serious security threat to Internet services in recent years. These attacks have evolved to be increasingly sophisticated, complex, and difficult to mitigate and detect. In this paper, we propose a new approach using HMM to detect DDoS attacks. The performance of the proposed approach is generally better and achieve higher detection rate and lower false positive rate comparing with two other machine-learning algorithms Naive Bayes and Neural Network. Training and testing applied on a DDoS data set with reduced feature. Using the reduced feature set after applying the Feature Pruning algorithm that we implemented obtains a significant improvement in detection performance and reduction model training and testing time.
关键词：Hidden Markov models (HMM); distributed denial of service (DDoS).
Loading...
