An Enhanced Anomaly- Based Model for Network Intrusion Detection Using Neural Network

Abstract

Anomaly-based IDSs need to be able to learn the dynamically changing behavior of users or systems. In this thesis, we are experimenting with packet behavior as parameters in anomaly intrusion detection. This research work has improved on existing Network Intrusion System and was motivated by the inability of some internet security to automatically prevent dangerous attacks. The developed IDS uses a back propagation artificial neural network (ANN) to learn system's behaviour. The methodology that was used for this research work is Object Oriented System Analysis Design and Methodology (OOADM), programming languages used are JavaScript for controls and flexibility, PHP for effective linking and communication with the database machine, HTML for browser communicator and MYSQL as a database machine. This research enhances the quality, convenience and reliability of Network Intrusion Detection System in internet services using artificial Neural Network, thereby providing a platform whereby information can be shared among internet users and in turn reduce the time spent by users in checking numerous intrusion attacks.