Using Kernel Density Estimator To Determine the Limits of Multivariate Control Charts.

Abstract

Quality control is an effective statistical tool in the field of controlling the productivity to monitor and confirm the manufactured products to the standard qualities and the certified criteria for some products and services and its main purpose is to cope with the production and industrial development in the business and competitive market. Quality control charts are used to monitor the qualitative properties of the production procedures in addition to detecting the abnormal deviations in the production procedure. The multivariate Kernel Density Estimator control charts method was used which is one of the nonparametric methods that doesn’t require any assumptions regarding the distribution of the data or determine the control limits in monitoring the production procedure when the data doesn’t follow the normal distribution or has an unknown distribution. The aim of this paper is to monitor the production procedure throughout a number of variables simultaneously to reflect the quality of the produced material. Simulation experiments were used with deferent significance levels to illustrate the way in which the multivariate Kernel Density Estimator charts method work while adopting the average range length criterion to demonstrate the performance and efficiency of the used methods. The results show that the method of nonparametric Kernel Density Estimator charts for the Gaussian Kernel had a good performance especially at significance levels with long-range