1. UKM Learning and Research Repository
2. Master Thesis / Tesis Master
3. Faculty of Information Science and Technology / Fakulti Teknologi dan Sains Maklumat

Title:	Hybrid bees algorithm for feature selection in network anomaly detection
Authors:	Osama Ahmad Alomari (P53666)
Supervisor:	Zulaiha Ali Othman, Assoc. Prof. Dr.
Keywords:	Intrusion detection systems (Computer security) Computer networks -- Security measures Anomaly detection (Computer security) Universiti Kebangsaan Malaysia -- Dissertations Dissertations, Academic -- Malaysia
Issue Date:	12-Jan-2012
Description:	In network security systems the intrusion detection (ID) is the most significant that detects several types of attacks. In today’s world of high volume of data transmission the need of high quality intrusion detection system (IDS) is increasing. Of late the intrusion detection has drawn the attention of a lot of researches, especially in data mining community. Generally the IDS deal with a huge volume of data traffic which contains irrelevant and redundant features hence it has become vital research topic. The quality of IDS is primarily influenced by feature selection. A lot of optimisation techniques are available for the purpose of feature selection such as: Genetic algorithm, Ant colony and Particle Swarm Optimisation. The Bees algorithm (BA) is relatively a new optimization technique that has been successfully applied in various optimisation applications such as job scheduling, data clustering, computer vision, and image analysis. Therefore, this research is aimed at investigating the performance of BA and hybrid bees algorithm with the simulating annealing (BSA) for feature selection to produce quality IDS. This research is conducted in two phases. In the first phase the basic BA is applied as feature selection technique and in the second phase BSA is applied. The feature selection techniques use a wrapper approach as a random search technique for subset generation and the Support Vector Machine (SVM) as the classifier technique. In this research the experiments were carried out using four random datasets collected from KDD-cup 99 and each dataset containing around 4000 records. The performance of the feature selection techniques are evaluated based on the detection rate and false alarm rate by comparing them with other feature selection techniques such as, Rough-DPSO, Rough, Markov_Blanket, Linear Genetic Programming (LGP), Support Vector Decision Function Ranking (SVDF) and Multivariate Regression Splines (MARS). The experiment result proved that the basic BA feature selection technique has produced better quality IDS as against the BSA. While the BSA has performed better, when compared with other techniques. Therefore this research concludes that the basic BA has become a good alternative feature selection technique for IDS.,“Certification of Master’s/Doctoral Thesis” is not available,Master Information Technology
Pages:	93
Call Number:	TK5105.59.A445 2012 3 tesis
Publisher:	UKM, Bangi
Appears in Collections:	Faculty of Information Science and Technology / Fakulti Teknologi dan Sains Maklumat

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.