Using the kurtosis correction method to design quality control limits for asymmetrical distributions: A comparative study in Kirkuk city

Abstract

In this paper, the limits of the mean and the range are constructed when the normal distribution condition is not available for the measured quality (relative humidity in Kirkuk city). By using kurtosis correction as well as Johnson's transform to convert the data to normal distribution, it is appeared to be subjected to the smallest extreme value distribution and then compared the new charts with Shewhart chart using process capability. The results show that the new charts are more efficient and sensitive to changes in the production process, which makes the proposed method of dealing with asymmetric data more benefit.