Level: Intermediate
Bradley Jones, JMP Distinguished Research Fellow, SAS
In this talk I will introduce a new method for constructing supersaturated designs (SSDs). The method leads to a partitioning of the columns of the design so that the columns within a group are correlated to the others within the same group, but are orthogonal to any factor in any other group. These new designs are called group-orthogonal supersaturated designs (GO-SSDs). Using this group structure it is possible to find an unbiased estimate of the error variance and develop an effective, design-based model selection procedure. Simulation results show that the use of these designs, in conjunction with our model selection procedure, enables the identification of larger numbers of active main effects than have previously been reported for supersaturated designs. These designs and their automated analysis are new in JMP 15. This talk will provide an example of this new design approach applied to the time it takes the custom design tool in JMP to create a design as a function of 12 factors. With only 12 runs, the design correctly identified the effects of the five most important factors.