Ensemble classifiers for network intrusion detection system

Abstract

Two of the major challenges in designing anomaly intrusion detection are to maximize detection accuracy and to minimize false alarm rate. In addressing this issue, this paper proposes an ensemble of one-class classifiers where each adopts different learning paradigms. The techniques deployed in this ensemble model are; Linear Genetic Programming (LGP), Adaptive Neural Fuzzy Inference System (ANFIS) and Random Forest (RF). The strengths from the individual models were evaluated and ensemble rule was formulated. Prior to classification, a 2-tier feature selection process was performed to expedite the detection process. Empirical results show an improvement in detection accuracy for all classes of network traffic; Normal, Probe, DoS, U2R and R2L. Random Forest, which is an ensemble learning technique that generates many classification trees and aggregates the individual result was also able to address imbalance dataset problem that many of machine learning techniques fail to sufficiently address it.