Some Statistical Control Charts On Autocorrelated Process

In our daily life, we need to identify individuals whose longitudinal behavior is different from the behavior of those well-functioning individuals so that some unpleasant consequences can be avoided.Advances in automated sampling technology have made autocorrelated data commonplace.Several statistical methods are used to reduce data autocorrelation and make statis-tics base on autocorrelated data independent.There are model-based and model-free meth-ods.We compare these methods in both Shewhart and cumulative sum(CUSUM)control charts using AR(2) model using the ANOS rule. We also use data from tobacco factory to do example verification and compare these three methods.Finally we conclude that UBM method is best, especially when autocorrelation is high. When batch size is relative big, the performance is better. The artical provides reference and direction for the practical application through the summary of simulation of three methods.