Multivariate zero-inflated Poisson models and their applications

摘要: The zero-inflated Poisson (ZIP) distribution has been shown to be useful for modeling outcomes of manufacturing processes producing numerous defect-free products. When there are several types defects, the multivariate ZIP (MZIP) model can detect specific process equipment problems and reduce multiple defects simultaneously. This article proposes MZIP models investigates distributional properties an model. Finite-sample simulation studies show that, compared method moments, maximum likelihood smaller bias variance, as well more accurate coverage probability in estimating parameters zero-defect probability. Real-life examples from a major electronic manufacturer illustrate how proposed procedures environment equipment-fault detection covariate effect studies.