%0 Journal Article
%T Intelligent Agent-Based Intrusion Detection System Using Enhanced Multiclass SVM
%A S. Ganapathy
%A P. Yogesh
%A A. Kannan
%J Computational Intelligence and Neuroscience
%D 2012
%I Hindawi Publishing Corporation
%R 10.1155/2012/850259
%X Intrusion detection systems were used in the past along with various techniques to detect intrusions in networks effectively. However, most of these systems are able to detect the intruders only with high false alarm rate. In this paper, we propose a new intelligent agent-based intrusion detection model for mobile ad hoc networks using a combination of attribute selection, outlier detection, and enhanced multiclass SVM classification methods. For this purpose, an effective preprocessing technique is proposed that improves the detection accuracy and reduces the processing time. Moreover, two new algorithms, namely, an Intelligent Agent Weighted Distance Outlier Detection algorithm and an Intelligent Agent-based Enhanced Multiclass Support Vector Machine algorithm are proposed for detecting the intruders in a distributed database environment that uses intelligent agents for trust management and coordination in transaction processing. The experimental results of the proposed model show that this system detects anomalies with low false alarm rate and high-detection rate when tested with KDD Cup 99 data set. 1. Introduction Mobile ad hoc networks (MANETs) consist of mobile nodes that work independently without an infrastructure. They are useful in application areas like disaster management emergency and rescue operations where it is not possible to have well-defined infrastructure. MANETs are characterized by its great flexibility. However, MANET¡¯s inherent vulnerability increases their security risks. Though MANETs are dynamic and cooperative in nature, it needs efficient and effective security mechanisms to safeguard the mobile nodes. Intrusion detection and prevention are primary mechanisms to reduce possible intrusions. Intrusion detection using classification algorithms effectively discriminates ¡°normal¡± behavior from ¡°abnormal¡± behavior. Therefore, intrusion detection and prevention system can be used as a secondary mechanism of defense in any wireless environment and mobile databases so that it can be a part of the reliable communication in MANETs [1]. Intrusion detection systems (IDS) play a major role in providing security to networks. In this paper, we introduce a new intelligent agent-based intrusion detection model for securing the mobile ad hoc networks. The main function of the proposed intrusion detection system is to monitor the computer system and network in order to find the intrusion activities in the system. In such system, attacks are divided into two categories, namely, host-based attacks and network-based attacks. Hence, IDSs are also
%U http://www.hindawi.com/journals/cin/2012/850259/