![]() | Only 14 pages are availabe for public view |
Abstract The most important use of a control chart is to improve the performance by identify assignable causes. If these causes can be eliminated from the process, variability will be reduced and the process will be improved, focused on the statistical technique using the regression analysis in constructing regression residual control chart. Regression analysis is one of the most commonly statistical techniques used for analyzing data, describing the relation between the response variable and the explanatory variables. Parametric regression is assume a pre-specified form of the regression function. One of the most common parametric estimation methods is maximum likelihood estimation that is used in analysis for fitting the models. On the other side, the nonparametric regression models do not assume a pre-specified form of the regression function, and one of the nonparametric estimation method which used in analysis is the spline smoothing. The design of control charts is performed through two phase{u2019}s analysis. In phase I analysis, the stable control chart for regression residual will be suggested for monitoring by using X-bar /S-charts. In phase II analysis, the monitoring done by using EWMA control charts. A simulation study has been conducted to compare between the performance of deviance residuals, and Pearson residuals in case of parametric regression estimation and nonparametric regression estimation through EWMA control charts. Applications have been done using R program version 3.5.2 for fitting a gamma model with two different link functions; identity link function and log link function and extracting two different types of residuals; deviance residuals and pearson residuals, then the average run length measure has been calculated to evaluate the performance of residual control charts |