A Futuristic Analysis Approach of Neural Network for Intrusion Detection System.

Abstract

Security is a crucial issue for computer systems; so, Intrusion Detection Systems began to exist as the next level of security. Artificial intelligence techniques provide a conceivable solution to intrusion detection. In this document, a qualitative approach of neural networks is proposed to detect intrusions based on analysis of relevant characteristics. The results show that separate neural networks based on the analysis of relevant characteristics for each type of attack detect attacks faster than the unique neural network classifier with the data set we have used from the KDD Cup data. The neural network approach reduces training time and attacks can be detected over time.