The performance of MEMWA control chart in generalized additive profile monitoring

The application of generalized additive models (GAMs) in profile monitoring is recently proposed. They seem to have satisfying results in monitoring complex processes, but still need more assessment in facing different situations. In this work, the performance of multivariate exponentially weighted average (MEWMA) control chart in generalized additive profiles is considered. MEWMA control charts are expected to detect small shifts if well adjusted. This behaviour is assessed in generalized additive profiles and is compared with Hotelling’s T 2 control chart. The analysis and comparisons are performed in monitoring monthly stroke types, including ischemic and hemorrhagic strokes of patients with acute stroke in Mashhad Stroke Incidence Study.

Keywords:

Generalized additive models , multivariate exponentially weighted moving average , multivariate control chart , profile monitoring , statistical process control.