An Effective Intrusion Detection System Based onMulti-layers Mining Methods

In this paper, we propose a multi-layer selection and mining methods for effective intrusion detection, which utilize feature selection, classification, clustering and evidence theory for decision making. In the experiments, DARPA KDD-99 intrusion detection data set is used for evaluation. It shows that our proposed classifier not only classifies and separates the normal and abnormal data, but also reduces false positive and false negative besides detecting all four kinds of attacks.