PCA mix‐based Hotelling's T ² multivariate control charts for intrusion detection system

Most of the data, which is in field network intrusion detection, have characteristics a mixture high-dimensional datasets continuous and categorical variables. It easily leads traditional multivariate control chart to get error detection results. Hotelling's T2 charts based on Principal Component Analysis mix (PCA mix) with bootstrap limit were proposed, applied system. was compared conventional PCA performance limits obtained method ones calculated using most commonly used kernel density estimation. The experimental results revealed that proposed had better than its counterparts.