A Bayesian Approach to Diagnosing General Linear Profiles

Apart from quick monitoring of abnormal changes in a multivariate process, it is also critical to accurately identify the cause of abnormal changes after a signal in multivariate statistical process control. Most diagnosis methods focus on the distribution of mass characteristics such as mean and/or variance. But the quality of a process may be better characterized by the relationship between the response variable and one or more explanatory variables in many applications, which is called profile problems in literatures. This paper develops a Bayesian approach to diagnosis parameter shifts in profile process. The proposed approach not only accurately identifies shift parameters but also provides the probabilities of shift parameters. Compared with existing methods, the proposed approach outperforms them.