Process Monitoring for Multiple Count Data Using a Deleted-YStatistic

Abstract

A procedure for monitoring discrete multivariate counts is proposed. The procedure is based on two schemes: the first scheme detects a shift in one or two of the p variables; the second procedure detects a shift in all the variables. The first scheme is based on an unbiased estimator for the shift in a particular Poisson count when input variables are not measurable. Run length results show that for most shift conditions this estimator is more effective in detecting a shift in one or two variables than using multiple C charts on the raw counts. The second scheme is based on the average of the p variables. Run length results show that its performance is equal to or better than the performance of multiple C charts on the raw counts. Together the two schemes provide efficient detection of virtually all shift situations.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 13 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.