A rule generation model using S-PSO for Misuse Intrusion Detection
Yi Zhang, Zhang Li-jun · 2010
Facing the increasingly important problem of computer security, Intrusion Detection System (IDS) has become an essential mechanism to protect computer and network system from malicious behaviors. In pursing high accuracy of detection rate, research in IDS is focusing on rule generation. Developing rules manually through human analysis on attack signatures often results in meaningful but costly work as it is difficult to define threshold. In this paper, we present a rule generation model for Misuse Intrusion Detection using a combination of statistical approach and particle swarm optimization (PSO) to achieve the rapid feature selection and rule optimization. Experimental results prove the effectiveness and robustness of the model we proposed, rules generated from which show both a high classification rate and a low false positive rate.