Design space regions do not match.To practice using JMP, I am analyzing and reproducing data from papers.
Among these, screening using the Plackett-Burman design and optimization using the central composite design were performed for three impurities to create a predictive model.
Similar predictive model equations were created for all of the models described in the paper.
The paper states that pooling was performed during the optimization process, but a model including all parameters was used to create a design space that combined the models.
Following this, we created a predictive model equation without pooling, i.e., including all parameters, to create a design space.
However, the area of the design space where impurities are 0.1% or greater does not match, and my design is wider.
The equation for the predictive model used to create the design space is not disclosed, so it is unclear, so what is the reason for this?
Sorry for the poor writing.
Design of ExperimentsRe: デザインスペースの領域が一致しません。Hi @keita_ ,
There are a few thoughts on this:
1. The authors have created the 'whole parameter' model and further refined it (i.e. with stepwise regression to reduce model terms) and then performed Design Space analysis - this would make more sense as you want to refine your model and reduce your non statistically significant terms before you attempt to perform Design Space profiling: if the model is overfitting then the profiles will be misleading.
2. Has the author mentioned any blocking considerations in the model? Have they discussed if they combined their Placket and central composite results together for the model creation?
3. The other (more unfortunate but common) likelihood is the results aren't reproducible, unfortunately this can happen in publications which is often why authors are reluctant to share their full data!
Hope this helps,
Many thanks,
Ben
Re: デザインスペースの領域が一致しません。Hi @Ben_BarrIngh ,
Thanks for the reply.
I am not very familiar with 2. blocking, but I don't think it was mentioned.
Also, regarding 3, the predictive model equation for design space is not disclosed.
It was very informative. Thank you.
Re: Design space regions do not match.My name is Ono, and I am a senior tester in the JMP Japan division of SAS Institute Japan.
It's been a while since the problem was solved, and my reply below may be off the mark, but I thought there might be a discrepancy in the question of how to create a rectangular area within an ellipse.
For example, suppose the area inside the red ellipse in the figure below is the area where the impurity rate is 0.1% or less. In that case, the rectangles (rectangular areas) that can be set as the design space are yellow and light blue.
In this example, if you want X1 to have width, the light blue rectangle is more practical, and if you want X2 to have width, the yellow one is more preferable. I feel that unless you add some kind of constraint, such as "Find a rectangle that is located inside the ellipse and has the largest possible area," the rectangle (rectangular region) cannot be uniquely determined.
