Automatic Intelligent Real-time Intrusion Detection System: A Data Mining Model
Automatic Intelligent Real-time Intrusion Detection System: A Data Mining Model
Abstract
Intrusion Detection System (IDS) is a popular device for securing network resources. However, developing a system that will automatically detect all known and unknown network attacks accurately in real-time is still a serious challenge. In this paper, a data mining model of an automatic intelligent real time network intrusion detection system that will be able to detect malicious network packets with very minimal false positive and false negative was presented. A three-layered Intrusion Detection System was modelled using Clustering techniques, Artificial Neural Networks and Decision Tree Classifiers. The algorithms used for each stage of the process were compared. Simple K-Means and Expectation Maximization Algorithms are compared in Layer One. In Layer Two, Multilayer Perceptron and Radial Basis Function Feed-Forward Back Propagation Neural Networks with varying parameters were compared while the best combination of the two was used. Equally, Simple Classification and Regression Tree, C4.5 and Naive Bayes Tree were tested in Layer Three. The model is used to develop architecture for a proposed Automatic Intelligent Real-time Intrusion Detection System (AIR-IDS).
Keywords
Intrusion Detection System; Data Mining; Intelligent System; Real Time;
Comments
Post a Comment