Block Bootstrap Control Limits for Multivariate Autocorrelated Processes
Kalgonda, A A
- Publisher: Journal of Engineering Computers & Applied Sciences
Journal of Engineering Computers & Applied Sciences
(issn: 2319-5606, eissn: 2319-5606)
Computer Science; Information Technology | Multivariate Control charts, Autocorrelation, Bootstrap
In production processes, due to automation the process can be sampled at higher rates, which leads autocorrelation. In practice, it is difficult to model such multivariate autocorrelated data adequately. If the underlying model for multivariate autocorrelated process is not adequate, or time series data possessing long memory property or when the form of the time series model is not exact the chart may fail to monitor the process correctly. Hence, it is essential to develop model free multivariate autocorrelated methods. In this article, control procedure based on Z-statistics applying bootstrap is used for monitoring an autocorrelated multivariate process. Example is given to illustrate the implementation of the procedure.