Performance of any control scheme in Phase II depends directly on the quality of estimators utilized in Phase I. In practice, outliers could be present in the data which would impact the performance of estimators adversely. This study deals with robust parameter estimation and monitoring linear profiles in the presence of outliers and compares the results with the least squares (LS) estimators. For this purpose, M-estimators are used as robust estimators and empirical distributions for related statistics are determined using Mont Carlo simulation to calculate control limits for two control charts and for codding independent variable method. Using a numerical example, profile parameters are estimated by ordinary least squares and M-estimators and the resulting statistics are monitored by two control schemes. Phase II control charts are determined based on the two types of estimators and compared for different out of control profiles. Empirical distributions did not follow their exact distributions obtained by least squares method. Simulation results confirm that M-estimators lead to better estimates in comparison to LS estimators and also improves classification performance. Robust estimators also lead to improvement in ARL performance in comparison to LS estimators.
9 September 2020 in Quality and Reliability Engineering International
Parameter estimation is the first step in constructing control charts. One of these parameters is the process mean. The classical estimators of the process mean are sensitive to the presence of outlying data and subgroups which contaminate the whole data. In existing robust estimators for the process mean, the effects of the presence of the individual outliers are being considered, while, in this paper, a robust estimator is being proposed to reduce the effect of outlying subgroups as well as the individual outliers within a subgroup. The proposed estimator was compared with some classical and robust estimators of the process mean. Although, its relative efficiency is fourth among the estimators tested, its robustness and efficiency are large when the outlying subgroups are present. Evaluation of the results indicated that the proposed estimator is less sensitive to the presence of outliers and the process mean performs well when there are no individual outliers or outlying subgroups.
7 October 2021 in Communications in Statistics - Theory and Methods
In recent years, some methods are proposed to analyze image data in statistical process control. These methods are mainly based on monitoring image data using statistical profile monitoring. In these methods, color images are converted to gray-scale images because of the simplicity of working. In this conversion, the color property of the images is generally removed, and the resulting profile does not reveal changes in the images. The main goal of this article is to develop a statistical approach based on multi-channel profiles for monitoring color image data. The proposed method applies a multi-channel functional principal component analysis to obtain a set of extracted features that can be effectively used to characterize process variations. These features can be used to construct an exponentially weighted moving average control chart. Numerous simulation studies are performed to evaluate the performance of the proposed method to detect shifts and change-point. Results for different fault sizes and smoothing constants indicate that the proposed method is capable of not only detecting shifts quickly but also estimating the change-point accurately. The results also illustrate the proper performance of the proposed approach in monitoring industrial cases to detect out-of-control conditions.