Detecting intrusions by using support vector machines
Dong Chun · Journal of Xidian University · 2003
The most common technology used in intrusion detection are anomaly detection technology and misuse detection technology, both of which are based on the mass prior knowledge. When given less prior knowledge, the performance of both techniques can not meet our need. Because SVM solves the small samples learning problem well, it will get good performance using SVM in intrusion detection, given less knowledge. A survey of intrusion detection is described and the basic concepts about SVM are introduced. The procedure of detection intrusions using SVM are discussed in detail. Finally, the experiment was performed on system call trace. The result shows that its performance is still good when given less prior knowledge.