What is the difference between a D-Optimal and I-Optimal designed experiment and when should you use one over the other?
Jun 18, 2019 5:43 AM
| Last Modified: Jun 20, 2019 12:59 PM
In his Basic Design of Experiments Mastering JMP session, Bill Worley @bill_worley suggests using D-Optimal if you want the best parameter estimates for the model. Use I-Optimal for the lowest prediction variance. If you think there is curvature, consider using I-Optimal, which will force mid-points into the design.