Construction of low false alarm and high precision RBFNN for detecting flooding based denial of service attacks using stochastic sensitivity measure

A good intrusion detection system (IDS) should have high precision on detecting attacks and low false alarm rates. Machine learning techniques for IDS usually yield high false alarm rate. In this work, we propose to construct host-based IDS for flooding-based denial of service (DoS) attacks by minimizing the generalization error bound of the IDS to reduce its false alarm rate and increase its precision. Experiments using artificial datasets support our claims.