A0599
Title: A rank-based EWMA chart for monitoring linear profiles when the random error is not normally distributed
Authors: Longcheen Huwang - National Tsing Hua University (Taiwan) [presenting]
Abstract: The profile monitoring of statistical process control is a technique for monitoring the stability of a functional relationship between a response variable and one or more variables over time. The monitoring of linear profiles is the most important one because the response variable and the covariate variables are usually correlated with linearity due to their flexibility and simplicity. Additionally, most of the existing charting methods for monitoring linear profiles assume that the random error is normally distributed. However, the normal assumption of the error term is not justified in certain applications. This causes the existing charting methods to be both inadequate and less efficient for monitoring linear profiles. Based on the rank-based regression, a charting scheme is developed for monitoring linear profiles where the random error is not assumed to be normal. The proposed charting scheme mainly applies the exponentially weighted moving average technique to the spatial rank of the vector of the Wilcoxon-type rank-based estimators of regression coefficients and an error variance estimator. Performance properties of the developed control chart are evaluated and compared with a multivariate sign charting scheme in terms of the out-of-control average run length. A real example is also used to illustrate the applicability and implementation of the proposed charting method.
