Intrusion detection using pattern recognition methods

This paper applied pattern recognition techniques to the intrusion detection field to detect abnormal user or host behaviors.On KDD99 data set,the information gain measurement had been used to select the related features.Two supervised methods,SVM (Support Vector Machine) as well as MNN (Muhi-layer Neural Network),and two unsupervised methods,Single- Linkage and K-Means had been implemented and evaluated.The results showed that those methods above have demonstrated good detection rate and low false positive rate,which is promising in intrusion detection field.