Abordagens Paramétricas e Não Paramétricas para Monitoramento de Parâmetro de Locação – Caso Univariado

Abstract

The Shewhart control chart is a powerful statistical tool in process control. The operation of these control charts is the periodic sampling off items produced. They are analyzed according to some characteristic of interest. The quality characteristic can be an attribute or a variable. The chart contains two horizontal lines, called upper and lower control limits. The width of the range between these limits is chosen so that, when the sampling point is within the control limits, it is considered that the process is operating under control. However, when a point occurs outside these limits, it is considered that the process is out of control, requiring management intervention for the process to operating again in statistical control conditions. The in-control performance of non-parametric individuals control charts based on kernel estimators are studied by simulation. Three different procedures are adopted for kernel estimator bandwidth selection. It turns out that the alternative control charts are robust against deviations from symmetry and perform reasonably well under normality of the observations.