期刊名称：International Journal of Innovative Research in Science, Engineering and Technology
印刷版ISSN：2347-6710
电子版ISSN：2319-8753
出版年度：2015
卷号：4
期号：7
页码：5588
DOI：10.15680/IJIRSET.2015.0407102
出版社：S&S Publications
摘要：Intrusion detection is an essential and important technique in research field. One of the main challengesin the security management of large-scale high-speed networks is the detection of suspicious anomalies in networktraffic patterns due to different kinds of network attack. We give attacks normally identified by intrusion detectionsystems. Differentiation can be done in existing intrusion detection methods and systems based on the underlyingcomputational methods used. Intrusion detection methods started appearing in the last few years. In this paper wepropose an Intrusion detection method using three different methods. These methods are K-means clustering, neurofuzzy models and C4.5 algorithm. We propose a three level framework for Intrusion detection. In First step k-meansclustering are used to create number of training subsets. In second step different neuro fuzzy models are traineddepending on the training subsets. At last we perform classification using C4.5 decision tree algorithm and findwhether data is anomalous or not.
关键词：k-means ; Fuzzy Neural Network; Intrusion Detection System ; Network Intrusion Detection System ;SVM ; C 4.5.
