Competing with High Quality Data: Concepts, Tools, and Techniques for Building a Successful Approach to Data Quality

Chapter 12 Design and Development of Multivariate Diagnostic Systems

12.0 Introduction

In previous chapters, we have described the importance of data quality and how it can be measured and improved with the critical data element (CDE) view. After ensuring high-quality and reliable data, the data should be used in a meaningful way to derive important insights. The methodology described in this chapter is useful in ensuring high information quality that will help up us to formulate important insights. We usually have to deal with the information based on more than one variable or CDE in order to draw insights, which will need to be used in relation to one another to make important decisions. Systems with more than one CDE or variable can be called multivariate systems. Examples of multivariate systems include medical diagnosis systems, client relationship mechanisms, fraud detection systems, and fire alarm sensor systems. This chapter describes the Mahalanobis-Taguchi Strategy (MTS) and its applicability to developing a multivariate diagnostic system with a measurement scale. The Mahalanobis distance (MD) is used to measure the distances in a multivariate system, and Taguchi's principles are used to measure accuracy of the system and identify important variables that are sufficient for the measurement system. This methodology is becoming increasingly popular, evidenced by the many case applications around the globe.

12.1 The Mahalanobis-Taguchi Strategy

The Mahalanobis-Taguchi ...