Intrusion detection system using bootstrap resampling approach of T² control chart based on successive difference covariance matrix

The multivariate control chart is one of SPC method that is often used in intrusion detection. The Hotelling’s T² control chart with Successive Difference Covariance Matrix (SDCM) is the robust method that can detect outliers in the process data for individual observation. This method will effective to be applied in Intrusion Detection System (IDS) because it can detect the anomaly or outliers in the network. The problem arise when the exact distribution of this method has not determined. Bootstrap is one of the nonparametric method that widely used to estimate the parameter without any distribution assumption applied to overcome the problem. In this research, the Hotelling’s T² control chart is improved using the SDCM while its control limits is calculated using bootstrap resampling method. The proposed method is applied in IDS and its performance is compared to the other control chart approaches. The performance evaluation result shows that the proposed IDS with bootstrap control limit performs better than the other control chart approaches for testing dataset. Moreover, the proposed IDS outperforms the other classification methods.