Hierarchical bayes quality measurement plan |
| |
Authors: | Kannan Natarajan Malay Ghosh Tapabrata Maiti |
| |
Affiliation: | 1. Bristol-Myers Squibb PRI , Princeton , NJ , 08543P.O. Box 4000;2. Department of Statistics , University of Florida , 103 Griffin-Floyd Hall, Gainesville , FL , 32611 |
| |
Abstract: | Quality Measurement Plan (QMP) as developed by Hoadley (1981) is a statistical method for analyzing discrete quality audit data which consist of the expected number of defects given the standard quality. The QMP is based on an empirical Bayes (EB) model of the audit sampling process. Despite its wide publicity, Hoadley's method has often been described as heuristic. In this paper we offer an hierarchical Bayes (HB) alternative to Hoadley's EB model, and overcome much of the criticism against this model. Gibbs sampling is used to implement the HB model proposed in this paper. Also, the convergence of the Gibbs sampler is monitored via the algorithm of Gelman and Rubin (1992). |
| |
Keywords: | audit sampling process empirical Bayes Gibbs sampling |
|
|