A Self-starting Control Chart for Monitoring Binary Response Profiles

In some processes the quality characteristic is characterizedby a functional relationship between a response variableand predictor variables which is named as profile. Theresponse variable in profiles usually follows normaldistribution. However, sometimes the response variableshave different distributions such as Bernoulli or Binomial.On the other hand, in many situations in real applicationsthe in-control (IC) parameters of the processes are notknown a priori and also there are no sufficient samples toexecute Phase I analysis to estimate the in-controlparameters. In this case, we desire to use self-startingcharting schemes that monitor the statues of processes. Inthis study, we consider logistic regression profiles whichhave binary response variable and propose a self-startingprocedure to monitor the process. The performance of theproposed method is evaluated by using average run length(ARL) criterion. The results showed a satisfactoryperformance of the proposed control chart.

Keywords:

Control chart , Self-starting , Profile monitoring , Binary response , Logistic regression profile , Average runlength (ARL)