An Effective Back Propagation Neural Network Architecture for the Development of an Efficient Anomaly Based Intrusion Detection System

The problem of intrusion is gradually becoming nightmare for several organizations. To protect the valuable data of their clients, organizations implement security systems to detect and prevent security breaches. But since the intruders are using sophisticated techniques to penetrate the systems, even the highly reputed secured systems have become vulnerable now. To deal with the current scenario, advanced level of researches are required to be carried out to invent more sophisticated Intrusion Detection System (IDS). Among various methodologies, researchers consider Back Propagation Neural Network (BPNN) as a very effective and popular tool for developing an IDS. In this paper, we have proposed an efficient BPNN architecture for the development of an anomaly based IDS with high accuracy and detection rate. The KDD´99 data set is used in this context to develop the architecture.