Meta learning intrusion detection in real time network
ICANN'07 Proceedings of the 17th international conference on Artificial neural networks
Hi-index | 0.00 |
To break the strong assumption that most of the trainingdata for intrusion detectors are readily available with highquality, conventional SVM, Robust SVM and one-class SVMare modified respectively in virtue of the idea from OnlineSupport Vector Machine (OSVM) in this paper, and theirperformances are compared with that of the original algorithms.Preliminary experiments with 1998 DARPA BSMdata set indicate that the modified SVMs can be trained onlineand the results outperform the original ones with lesssupport vectors(SVs) and training time without decreasingdetection accuracy. Both of these achievements benefit aneffective online intrusion detection system significantly.