Analysis of Ordered Categorical Data Using Expected Loss Minimization

Robust design is a strong method for the quality improvement of processes (or products). Two main techniques for data analysis in robust design process are the S/N ratio technique and accumulation analysis developed by Taguchi. Accumulation analysis is developed for the study of ordered categorical data. In this article, the current techniques are studied. Then, a new method that is more accurate than other methods is developed and introduced. In the next phase, an example from literature solved by the methods and the new method's optimality against other methods is shown.

Keywords:

• Quality engineering
• Robust design
• Experimental design
• Polysilicon
• VLSI

Notes

*Empty columns are identified by e.