Title: A comparison study and discussion of the Mahalanobis-Taguchi System

Authors: Seoung Bum Kim, Kwok-Leung Tsui, Thuntee Sukchotrat, Victoria C.P. Chen

Addresses: Department of Industrial and Manufacturing Systems Engineering, University of Texas at Arlington, Arlington, TX 76019, USA. ' School of Industrial and Systems Engineering, Georgia Institute of Technology, Atlanta, GA 30332, USA. ' Department of Industrial and Manufacturing Systems Engineering, University of Texas at Arlington, Arlington, TX 76019, USA. ' Department of Industrial and Manufacturing Systems Engineering, University of Texas at Arlington, Arlington, TX 76019, USA

Abstract: The Mahalanobis-Taguchi System (MTS) is a diagnostic tool for selecting the appropriate variables and quantifying the degree of abnormality in the group that needs to be diagnosed. MTS is said to be |data analytic| rather than |probability distribution based|. This statement has brought out some controversy in the statistics community. Despite the simplicity of MTS, it is considered ad hoc owing to the lack of statistical and probability theory. In this paper, we attempt to incorporate statistical thinking into the ad hoc parts of MTS by comparing with other multivariate statistical methods.

Keywords: Hotelling; T-1; loss functions; MTS; Mahalanobis-Taguchi system; PCA; principal component analysis; tree-based methods; multivariate statistical methods.

DOI: 10.1504/IJISE.2009.026768

International Journal of Industrial and Systems Engineering, 2009 Vol.4 No.6, pp.631 - 644

Published online: 26 Jun 2009 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article