Research On Construction And Application Of Robust Autocorrelation Control Chart

In Six Sigma management,the construction of the upper and lower limits of the conventional quality control chart is mainly based on the assumption that the controlled process is independent of the same distribution(that is,to satisfy independence,equal variance,and same distribution).But in recent years,with the rapid development of information technology,people's ability to collect data has become stronger and stronger,and they can record a large amount of product quality data quickly and accurately in the production process.Although a large amount of data is helpful to understand the quality of the product,there may also be outliers,autocorrelation and other abnormal scenes in the large amount of data,which interfere with the results of data analysis,and raise the flexibility and effectiveness of statistical methods.Challenge.Aiming at the complex features such as autocorrelation,outliers and different distributions(both outliers and autocorrelation)in the quality control chart,this paper proposes to introduce the idea of robust estimation into the construction of the quality control chart.That is,the original data with outliers are processed with a weighting function,and the original sample data is assigned different weights according to different residual values.Large errors are assigned small weights,and small errors are assigned large weights.Through robust weighting processing Reduce the effect of outliers on the upper and lower limits of the control chart.The results show that(1)no matter whether the controlled data is in a normal distribution or autocorrelation,the existence of outliers will greatly widen the upper and lower limits of the control chart and seriously weaken the control effect of the control chart;(2)comprehensively compare the current There are commonly used robust mean and standard deviation robust estimation methods,and it is found that the weighting function can more effectively reduce the influence of outliers on the control process and improve the control effect of the control chart;(3)Including both outliers and self When constructing a control chart based on the data of these two abnormal phenomena,the residual control chart processed by the weighting function can well resist the interference of outliers.