Real-Time Intelligent Multilayer Attack Classification System
Journal: ICTACT Journal on Soft Computing (IJSC) (Vol.4, No. 2)Publication Date: 2014-01-01
Authors : T Subbulakshmi S G Keerthiga; R Dharini;
Page : 677-686
Keywords : Distributed Denial Service of Attacks; Intrusion Detection System; Support Vector Machine; Neural Networks; Fuzzy Inference System;
Abstract
Intrusion Detection Systems (IDS) takes the lion’s share of the current security infrastructure. Detection of intrusions is vital for initiating the defensive procedures. Intrusion detection was done by statistical and distance based methods. A threshold value is used in these methods to indicate the level of normalcy. When the network traffic crosses the level of normalcy then above which it is flagged as anomalous. When there are occurrences of new intrusion events which are increasingly a key part of system security, the statistical techniques cannot detect them. To overcome this issue, learning techniques are used which helps in identifying new intrusion activities in a computer system. The objective of the proposed system designed in this paper is to classify the intrusions using an Intelligent Multi Layered Attack Classification System (IMLACS) which helps in detecting and classifying the intrusions with improved classification accuracy. The intelligent multi layered approach contains three intelligent layers. The first layer involves Binary Support Vector Machine classification for detecting the normal and attack. The second layer involves neural network classification to classify the attacks into classes of attacks. The third layer involves fuzzy inference system to classify the attacks into various subclasses. The proposed IMLACS can be able to detect an intrusion behavior of the networks since the system contains a three intelligent layer classification and better set of rules. Feature selection is also used to improve the time of detection. The experimental results show that the IMLACS achieves the Classification Rate of 97.31%.
Other Latest Articles
- PATH LOSS PREDICTION FOR GSM MOBILE NETWORKS FOR URBAN REGION OF ABA, SOUTH-EAST NIGERIA?
- The system approach and development of cybernetic representations
- Dynamic model behavior of the same market destination goods
- Application of games with incomplete information for the modellign of dynamic of price
- Conceptual model of coordination of regional investment project management processes
Last modified: 2014-02-18 18:11:49