Review of Machine Learning-Based Intrusion Detection Techniques for MANETs

Mobile ad hoc network is a widely developing technology that has been used in various areas such as in health care, military, virtual classrooms and conferences. However, mobile ad hoc networks are installed in critical situations security in this network is an important issue. Many susceptible characteristics of mobile ad hoc networks make an attacker breach the system easily. So, it is important to have an intrusion detection system which can monitor mobile ad hoc networks constantly to identify any suspicious behaviour. Anomaly and misuse detection are the two widely used intrusion detection mechanisms used to analyse the attacks in mobile ad hoc networks. Anomaly intrusion detectors were proven to be more effective against unknown attacks. A number of anomaly-based intrusion detectors based on machine learning techniques were developed and tested against various attacks. In this paper, several intrusion detection techniques which used machine learning approaches for detection are reviewed and a hybrid IDS technique which combines with genetic algorithm and Bayesian game theory is proposed.