Increasing performance Of intrusion detection system using neural network

Abstract

Rapid growth in Internet and in parallel attacks, vulnerability and threats, has made intrusion detection systems very essential component in all parts of security infrastructure. Building IDS is not a new task, classical signature based IDS are used but they are unable to handle novel attacks. In this paper artificial neural network based intrusion detection is proposed for complete KDD cup 99 dataset. Performance of the proposed ANN based IDS system is evaluated and results shows high anomaly detection accuracy for the complete KDD cup 99 dataset as compared to existing techniques.