Title: Intercept model errors and design of high sigma cause selecting control charts
Authors: S. Lakshminarasimhan, S.M. Kannan
Addresses: TVS Sewing Needles Limited, Madurai 625016, India. ' Department of Mechanical Engineering, Kalasalingam University, Krishnankoil 626190, India
Abstract: In pedagogy, regression models assume an intercept term. But in real industry situations, there may be instances depicted by a no intercept model. In this work, the influence of intercept model errors in the design of high sigma cause selecting control charts for bivariate cascade processes have been discussed. A modified design flow has been proposed. It is also established that, for this type of control chart, when an intercept model is assumed as no intercept model the type II error or probability of missed signals is deflated. In the case of a no intercept model, assumed as an intercept model, the type II error is inflated.
Keywords: ARL; average run length; cause selecting control charts; confidence interval; high sigma process; intercept model errors; missed signals; MSRes; mean square of residuals; type II error; process management; benchmarking; bivariate cascade processes; statistical process control; SPC.
DOI: 10.1504/IJPMB.2009.026409
International Journal of Process Management and Benchmarking, 2009 Vol.3 No.1, pp.60 - 75
Published online: 10 Jun 2009 *
Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article