An Empirical Depth-Based Statistical Process Control Scheme for Multivariate Data

• Activity Number: 430
• Type: Contributed
• Date/Time: Tuesday, August 11, 2015 : 2:00 PM to 3:50 PM
• Sponsor: Quality and Productivity Section
• Author(s): Junvie Pailden*
• Companies: Southern Illinois University Edwardsville
• Keywords: control chart ; empirical likelihoood ; ewma ; nonparametric ; average run length ; sequential rank

Abstract

In this talk, we introduce a nonparametric control chart of the multivariate process mean based on empirical depth measures. This empirical depth is defined through the empirical distribution of the sample and has continuous and convex contours which capture the shape of the data points. We also introduce an EWMA chart based on sequential ranks of the empirical depth measures. These control charts require minimal distributional assumptions and is ideal for monitoring individual multivariate Phase I observations. A limited simulation study is presented to compare the proposed control charts to traditional control charts under non-standard data distributions.