Semi-parametric Profile Monitoring Based On EWMA

In recent years,SPC methods are frequently applied to monitoring of the profile,and the profile is usually modeled using linear,non-linear and non-parametric models.Based on the semi-parametric model,this paper proposes a local mode regression methods to estimate the model.The MNEWMA control graph is established based on the generalized likelihood ratio test theory,and compared with the PNEWMA control graph which is based on local polynomial method.Both methods can effectively solve the monitoring problem under the semi-parametric model and over-come the limitation of parametric model.In this paper,through the numerical simulation model comparing the two control charts and the method under non parametric model,it is concluded that MNEWNA and PNEWMA control chart,are more effective than NEWMA control chart be-cause they can find out of control condition more timely.Besides,the MNEWMA faster and more accurate than PNEWMA judging the change point location.But PNEWMA is more stable than MNEWMA in anti-interference effect of random factors.At last,through the actual air quality data of the two parties method comparison,again to verify the efficiency of the two methods in the paper.