A Practical Approach for Removing Multiple Outliers Using the T-Square Statistic

• Activity Number: 539
• Type: Contributed
• Date/Time: Wednesday, August 4, 2010 : 10:30 AM to 12:20 PM
• Sponsor: Section on Quality and Productivity
• Author(s): John C. Young and Robert L. Mason*+ and Youn-Min Amanda Chou
• Companies: Retired and Southwest Research Institute and The University of Texas at San Antonio
• Keywords: Multivariate Statistical Process Control ; Masking ; Swamping

Abstract

The occurrence of certain patterns of multiple outliers may produce problems that weaken the ability of the T-square statistic to detect potential outliers in a Phase I operation. The most prominent of these are the problems of masking and swamping. A number of robust procedures have been developed as solutions to these two problems. However, the application of the majority of these procedures to a multivariate process is not easily accomplished. We propose a procedure based on the T-square statistic that addresses both the masking and swamping problems, and is easy to apply. Industrial data are used to illustrate the proposed multiple outlier detection procedure.