期刊名称：International Journal of Innovative Research in Computer and Communication Engineering
印刷版ISSN：2320-9798
电子版ISSN：2320-9801
出版年度：2016
卷号：4
期号：3
页码：4456
DOI：10.15680/IJIRCCE.2016.0403242
出版社：S&S Publications
摘要：Intrusion Detection System (IDS) is a system that used for detecting the misuse of system i.e. malicious attacks. This system works as software for security management. Different techniques have been proposed by many researchers for Intrusion Detection System to achieve the better accuracy. In this paper we proposed hybrid approach to intrusion detection system. This hybrid approach make uses of k-means clustering and neural network Multi-Layer Perceptron (MLP) classification, this help to improve performance of the system. In this system we used KDD cup'99 dataset.In this paper, we used clustering to divide the huge amount of data into different category. We have used k-means clustering for this purpose. In neural networks a set of nodes are used. These nodes are further used to label data for intrusion detection. In this way it gives solution to some of the neural networks inherent problems for classificationto overcome like as the slow speed while labeling. Also the overhead of convergence and the burden of computation while classifying.
