Modeling intrusion detection system using hybrid intelligent systems

The process of monitoring the events occurring in a computer system or network and analyzing them for sign of intrusions is known as intrusion detection system (IDS). This paper presents two hybrid approaches for modeling IDS. Decision trees (DT) and support vector machines (SVM) are combined as a hierarchical hybrid intelligent system model (DT–SVM) and an ensemble approach combining the base classifiers. The hybrid intrusion detection model combines the individual base classifiers and other hybrid machine learning paradigms to maximize detection accuracy and minimize computational complexity. Empirical results illustrate that the proposed hybrid systems provide more accurate intrusion detection systems. r 2005 Elsevier Ltd. All rights reserved.