Nonparametric Control Charts for Monitoring Serial Dependence based on Ordinal Patterns

We consider the problem of monitoring for existence serial dependence in real-valued and continuously distributed processes. These can exist when a process goes out-of-control. Besides, as many control charts are designed under assumption independent identically observations, validity this needs to be checked. In literature, majority studies handled by using specific case autoregressive moving-average time series models. Also moving-window approach is often considered, which leads drawback that at least window length observations need collected before an out-of-control situation detected. Here, we use ordinal patterns propose fully nonparametric distribution-free, have unique chart design, used almost instantaneously start monitoring. The proposed do not require any model fitting, thus, eliminating problems associated with estimation errors, but might part Phase-I study. Through comprehensive performance comparison study various scenarios, effectiveness uncovering shown recommendations selection given. Implementations illustrated batch yield data from chemical process.