Philip J. Ramsey, PhD, Owner, North Haven Group, and Professor, University of New Hampshire
Mia L. Stephens, JMP Academic Ambassador, SAS

Jones and Nachtscheim (2011) introduced a new type of highly efficient experimental design called definitive screening designs DSDs). The designs are unique in that all factors in the experiment have three levels (allowing for the potential estimation of quadratic effects), main effects are orthogonal and free of aliasing, and no two-factor interaction or quadratic effect is fully confounded with another effect; all of this in as little as 2K+1 trials for K factors. Although the DSDs are starting to be adopted in industry, especially in biotechnology, they do present the experimenters with unique opportunities and challenges in terms of analysis. In this talk we will discuss various modeling strategies for DSDs that can be easily implemented in JMP 11, and will provide several examples from biotechnology experiments. The methods discussed include: prediction averaging over a subset of models selected using AICc and BIC; model averaging to estimate a full quadratic model; a new technique we refer to as Pareto-based modeling; and the LASSO, a penalized regression technique in JMP Pro 11. All methods will be compared using models built on actual examples, and recommendations on the use of the various methods will be discussed.


7282_pastedImage_1.png

Published on ‎03-24-2025 09:02 AM by Community Manager Community Manager | Updated on ‎03-27-2025 09:49 AM

Philip J. Ramsey, PhD, Owner, North Haven Group, and Professor, University of New Hampshire
Mia L. Stephens, JMP Academic Ambassador, SAS

Jones and Nachtscheim (2011) introduced a new type of highly efficient experimental design called definitive screening designs DSDs). The designs are unique in that all factors in the experiment have three levels (allowing for the potential estimation of quadratic effects), main effects are orthogonal and free of aliasing, and no two-factor interaction or quadratic effect is fully confounded with another effect; all of this in as little as 2K+1 trials for K factors. Although the DSDs are starting to be adopted in industry, especially in biotechnology, they do present the experimenters with unique opportunities and challenges in terms of analysis. In this talk we will discuss various modeling strategies for DSDs that can be easily implemented in JMP 11, and will provide several examples from biotechnology experiments. The methods discussed include: prediction averaging over a subset of models selected using AICc and BIC; model averaging to estimate a full quadratic model; a new technique we refer to as Pareto-based modeling; and the LASSO, a penalized regression technique in JMP Pro 11. All methods will be compared using models built on actual examples, and recommendations on the use of the various methods will be discussed.


7282_pastedImage_1.png



Start:
Mon, Sep 15, 2014 09:00 AM EDT
End:
Fri, Sep 18, 2015 05:00 PM EDT
Attachments
0 Kudos