Network Intrusion Detection System Using Random Projection Technique

摘要

In this paper, we propose a novel and practical intrusion detection system using random projection technique, which can be used in the situation of very high dimensional dataset. To the best of our knowledge, this is the first time applying random projection technique into the intrusion detection areas. The main idea of our proposed intrusion detection system is to use random projection technique projecting the high-dimensional dataset to a lower dimension, and then do the intrusion detection. We present a methodology of using a Support Vector Machine (SVM) classifier to build the intrusion detection model. The approach is applied to a set of benchmark data KDDCup99, and the experimental results show that the proposed intrusion detection system can do the intrusion detection efficiently with less computational complexity and system resource requirements.