A Real-Time Anomaly Network Intrusion Detection System with High Accuracy

Authors

Ahmed A. Elngar, Scince faculty, Al-Azhar University, Cairo, EgyptFollow
Dowlat A. El A. Mohamed, Math & Computer Science Department, Faculty of Science, Ain-Shams UniversityFollow
Fayed F. M. Ghaleb, Math & Computer Science Department, Faculty of Science, Ain-Shams UniversityFollow

Abstract

Reliance on Internet and online procedures increased the potential of attacks launched over the Internet. Therefore, network security needs to be concerned to provide secure information channels. Intrusion Detection System (IDS) is a valuable tool for the defense-in-depth of computer networks. However, building an efficient IDS faces a number of challenges. One of the important challenges is dealing with data containing high number of features. This paper is devoted to solve this challenge by proposing an effective PSO-Discritize-HNB intrusion detection system. The proposed PSO-Discritize-HNB IDS combines Particle Swarm Optimization (PSO) and Information Entropy Minimization (IEM) discritize method with the Hidden Naive Bayes (HNB) classifier. To evaluate the performance of the proposed network IDS several experiments are conducted on the NSL-KDD network intrusion detection dataset. A comparative study of applying Information Gain (IG) which is a well known feature selection algorithm with HNB classifier was accomplished. Also, to validate the proposed PSO-Discritize-HNB network intrusion detection; it is compared with different feature selection methods as Principal Component Analysis (PCA) and Gain Ratio. The results obtained showed the adequacy of the proposed network IDS by reducing the number of features from 41 to 11, which leads to high intrusion detection accuracy (98.2%) and improving the speed to 0.18 sec.

Recommended Citation

A. Elngar, Ahmed; A. El A. Mohamed, Dowlat; and F. M. Ghaleb, Fayed (2013) "A Real-Time Anomaly Network Intrusion Detection System with High Accuracy," Information Sciences Letters: Vol. 2 : Iss. 2 , PP -.
Available at: https://digitalcommons.aaru.edu.jo/isl/vol2/iss2/1