Profile monitoring for a binary response

Pertaining to industrial applications in which the response variable of interest is binary, this paper studies how the profile functional relationship between the response and predictor variables can be monitored using logistic regression. Under such a premise, several Hotelling T-2 charts that have been studied under continuous response variable to binary response variable for the purpose of Phase I profile monitoring are extended. The performance of these T-2 charts in terms of the signal probability for different out-of-control scenarios is compared based on simulation studies. A real example originated from aircraft construction is given in which these T-2 charts are applied and compared using the data. A discussion of potential future research is also given.