ABSTRACT
In some applications, the quality of a process or product is characterized by correlated multivariate linear and generalized linear model (GLM) regression profiles. Monitoring these profiles separately leads to misleading results because the correlation structure among the multivariate linear and GLM profiles is neglected. In this paper, we specifically concentrate on Phase II and propose some procedures for monitoring multivariate linear and GLM regression profiles. Simulation studies are used to compare the performance of the proposed methods under different magnitudes of shifts in the regression parameters in terms of the average run length criterion. The results of simulation studies show the superior performance of the proposed methods compared to monitoring multivariate linear and GLM profiles separately. In addition, the performance of the proposed monitoring schemes is illustrated by a numerical example. Finally, the application of the proposed methods is shown by a real-world case.
Acknowledgements
The authors are thankful to the respectful referees for precious comments which led to significant improvement in quality of the paper. The authors also appreciate Prof Guido Masarotto, Prof Giovanna Capizzi and Prof Philippe Castagliola, the guest editors of QTQM/ISSPM 2015 special issue for their precious comments as well as handling the paper.