Modeling a self-learning detection engine automatically for IDS

Intrusion detection systems (IDS) have become important and widely used tools for ensuring network security. Most IDS have previously been built by hand and they have difficulty in successfully classifying intruders because they need a significant amount of intrusion signatures. This paper describes a new IDS modeling method that uses machine-learning technology to automatically model a detection engine (DE) and has the ability to boost its performance using unlabeled data by means of automatic self-learning.