Advances in modern data acquisition techniques and computing power have enabled the collection and analysis of many quality characteristics simultaneously. If these quality characteristics are monitored separately, it might not be very effective in detecting process changes. To simultaneously monitor multiple characteristics, multivariate charts are being used. Though many studies are available in this field, they mostly assume multivariate normal distributions. This may be true in manufacturing industries but need not necessarily hold good in service sector. N such situations traditional multivariate approaches may lead to misleading conclusions. There are studies that proposed generalized Poisson distribution for multivariate count data. A drawback of this approach is that they do not allow zero or negative correlation. Though there are some approaches to overcome these drawbacks were proposed they lack in systematic methodology to monitor the multivariate count data. To overcome this problem, this study proposes a multivariate exponentially weighted moving average (MEWMA) scheme to monitor multivariate count data applying generalized Poisson distribution. (43 refs.)
展开▼
