346172_en-US

liliconcom

GO SSD

Has anyone either cooked up or found a good example for GO SSD?

 

I went through Brad's example that is on the technical reference library.  It's useful for me but probably not for my customer. I need to demo for a manufacturing customer so looking for something with manufacturing variables that may already have a dataset associated with it from which I can simulate responses. I have tried to simulate something based on white polymer. Initial attempts aren't so good and I think it will take some time to make a synthetic example work. If anyone has something I would be most grateful if you could share.

Re: GO SSD

Hi Jason,

 

Sure, sounds great, I'd like to see all of it.  I can include it in the JMP Discovered DOE project.

 

Hadley

Re: GO SSD

Hey Hadley,

 

I will upload my table and a script that launches and populates the GO SSD designer after I do a few refinements. Would you be willing to look at it? I would love to get your feedback and suggestions for updates. I also have a DOE project that uses white polymer for custom, DSD and now GO SSD examples. I am using this in lieu of a journal. I am happy to upload this as well. 

 

J 

Re: GO SSD

Hi Jason,

 

Thanks for taking the initiative on this.  Would you be able to share the table or journal with us?

Re: GO SSD

Operator error after all. I failed to add noise to the response simulation formula in the actual factor level table. Not sure how many times I looked at that and didn't see it! Thanks once again @karencopeland  it was fun working through this with you!

 

Everything is working as it should. The results I am getting are solid and I am stoked to demo this next week.

Re: GO SSD

Thank you, Jeff!

Re: GO SSD

Always check the red triangle menu:

 

 

Re: GO SSD

Still having issues. Is there a way to edit my previous post? I have not found the solution yet. 

Re: GO SSD

Originally wrote: Okay. Operator error here. Actual factor levels are in a Coding column property. My simulation formula that works on the -1, 1 coding works for the table with actual factor level coding. Thank you Karen Copeland for helping me puzzle that out.

 

Edited statement: While simulation works, Fit GO SSD does not. Fit GO SSD on table with actual factor levels shows no active factors. In the same table if I fit a model with known active factors in Fit Model I also get some very weird results. This is puzzling.  If I use a table with with -1, 1 coding everything works fine. The report on the left below is from the table with actual factor coding. The one on the right is with -1,1 coding. 

 

 

Re: GO SSD

I am close to having a workable data table to demo based on white polymer. I found I could not use actual factor levels when using Fit GO SSD. With coded factor levels all seems to work. Am I missing something or is the platform limited to coded variables? I have read through the documentation and JMP 15 training starter and did not find an explanation. I can see how it makes sense for the analysis but seeing actual factor levels is more interpretable and useful when running the experiment.