A Hybrid Intrusion Detection System Based on Multilayer Artificial Neural Network and Intelligent Feature Selection

Abstract

Abstract. Increased intrusions into computer networks and cyber-attacks have rendered the immunization of cyberspace one of the most important issues of managers and experts in the recent years. Since cyber-attacks have become more sophisticated and hackers have become more professional, mere use techniques such as firewall, cryptography, biometrics, and antiviruses is not sufficient anymore. Therefore, it is necessary to employ efficient intrusion detection systems. Considering 5 classes of cyber-attacks, a detection intrusion system, of abuse detection type, based on the combination of a multilayer artificial neural network and an intelligent feature selection method was introduced in this research. The research results indicated that the feature selection phase using the proposed method yielded more favorable outcomes than the compared method in terms of the evaluation criteria.