Enhancement and evaluation of pattern recognition in control charts
International Journal of Quality & Reliability Management
ISSN: 0265-671X
Article publication date: 1 April 2001
Abstract
Pattern recognition applied to control charts centers around the development and assessment of automated algorithms for detecting non‐random or unnatural patterns in observations collected from a production process. The work presented here marks the first examination of enhancements to an existing algorithm, of investigations into sensitivity analysis issues, of development of standard performance metrics, and of a comparative performance with the traditional Western Electric Run tests. The simulation results of the research presented here indicate that the modified algorithm performs markedly better than the original algorithm, is only slightly sensitive to the selection of the user specified algorithm parameters, and competes favorably with the Western Electric Run Tests especially when detecting repetitive patterns like cycles.
Keywords
Citation
Piplani, R. and Faris Hubele, N. (2001), "Enhancement and evaluation of pattern recognition in control charts", International Journal of Quality & Reliability Management, Vol. 18 No. 3, pp. 237-253. https://doi.org/10.1108/02656710110383511
Publisher
:MCB UP Ltd
Copyright © 2001, MCB UP Limited