Application of Multivariate Statistical Analysis in Process Control

Abstract

Due to significant increase of information systems and its intensive usage in our everyday life, several problems like automatic identification of system faults, finding times of drastic change in stochastic characteristics as well as locating those characteristics, which “went out of control” need to be addressed. To solve these problems, we propose an algorithm based on multivariate statistical analysis. The algorithm is implemented with the R software environment and tested on custom metrics for Vesta server and other groups of random metrics.