Anomaly Detection for Compositional Data using VSI MEWMA control chart

In recent years, the monitoring of compositional data using control charts has been investigated in Statistical Process Control field. this study, we will design a Phase II Multivariate Exponentially Weighted Moving Average (MEWMA) chart with variable sampling intervals to monitor based on isometric log-ratio transformation. The Time Signal be computed Markov chain approach investigate performance proposed chart. We also propose an optimization procedure obtain optimal limit, smoothing constant, and out-of-control for different shift sizes short intervals. comparison standard MEWMA is provided. Finally, end paper conclusion some recommendations future research.