Adaboost Ensemble with Genetic Algorithm Post Optimization for Intrusion Detection
Abstract
This paper presents a fast learning algorithm using Adaboost ensemble with simple genetic algorithms (GAs) for intrusion detection systems. Unlike traditional approaches using Adaboost algorithms, it proposed a Genetic Algorithm post optimization procedure for the found classifiers and their coefficients removing the redundancy classifiers which cause higher error rates and leading to shorter final classifiers and a speedup of classification. This approach has been implemented and tested on the NSL-KDD dataset and its experimental results show that the method reduces the complexity of computation, while maintaining the high detection accuracy. Moreover, the method improves the processing time, so it is especially appealing for the real-time processing of the intrusion detection system.
Keywords: Intrusion Detection; AdaBoost; Genetic Algorithm; Feature Selection; Classification; NSL-KDD dataset.
Download Full-Text
ABOUT THE AUTHORS
Hany M. Harb
Computers and Systems Engineering Dept., Faculty of Eng., Azhar University
Abeer S. Desuky
Mathematics Dept., Faculty of Science, Azhar University
Hany M. Harb
Computers and Systems Engineering Dept., Faculty of Eng., Azhar University
Abeer S. Desuky
Mathematics Dept., Faculty of Science, Azhar University