Improving Taguchi's Packaging of Fractional Factorial Designs
G. K. Robinson
Abstract
G. K. Robinson
Abstract
Two changes to Taguchi's packaging of fractional factorial designs are suggested. First, Tsui's idea of confounding tables is supported. Second, a modification of Taguchi's linear graphs is proposed in which the way vertices and lines are drawn indicates whether the main effects and interactions corresponding to those vertices and lines are confounded with other two-factor interactions. This modification tends to encourage the use of a different set of linear graphs and thereby to encourage the use of higher resolution designs.
OpenAlex reports 7 citations for this work. Citation counts describe recorded attention and do not establish research quality.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
Two changes to Taguchi's packaging of fractional factorial designs are suggested. First, Tsui's idea of confounding tables is supported. Second, a modification of Taguchi's linear graphs is proposed in which the way vertices and lines are drawn indicates whether the main effects and interactions corresponding to those vertices and lines are confounded with other two-factor interactions. This modification tends to encourage the use of a different set of linear graphs and thereby to encourage the use of higher resolution designs.
Key concepts: Taguchi methods, Fractional factorial design, Factorial experiment, Mathematics, Orthogonal array, Combinatorics, Set (abstract data type), Statistics