RFAODE: A Novel Ensemble Intrusion Detection System

In recent years information and communication technology (ICT) has become an important part of human life. But ICT brings a lot of cyber risks. New threats and vulnerabilities are created to attack network system. Intrusion detection system (IDS) is used to detect these attacks. Machine learning (ML) and Data Mining (DM) techniques are widely used for IDS. Current IDS algorithms result in high error rate and less accurate to classify various attacks. This paper deals with a novel ensemble classifier (RFAODE) for intrusion detection system. Proposed ensemble classifier is built using two well-known algorithms RF and AODE. Average One-Dependence Estimator (AODE) resolved the attribute dependency issue in Naïve bayes classifier. Random Forest (RF) improves accuracy and reduces the error rate. The performance of proposed ensemble classifier (RF+AODE) is analyzed on Kyoto data set. With accuracy of 90.51% and FAR of 0.14, proposed ensemble classifier outperforms AODE, Naïve bayes, and RF algorithms and efficiently classifies the network traffic as normal or malicious.