topic Re: How do I get a list of all combinations that can give greater than a threshold of the response in JMP DoE RSM model in Discussions

How do I get a list of all combinations that can give greater than a threshold of the response in JMP DoE RSM model

Do_AuLu — Mon, 19 Feb 2024 12:20:07 GMT

This may sound like a basic question. I am a fairly new user.

I had designed a RSM model with four factors A,B, C, and D. After conducting the experiments and loading the results, the model suggests, all three factors are significant and that the interaction factor between A and D is also significant- with positive correlation. The interaction between factors B and C is a significant factor but has a negative correlation.

My aim is to find the lowest amount of each excipient that can give me a response greater than 15. I have used prediction profiler to identify concentration of each factor that gives me this response when other excipients are kept minimum. Due to the the significance of the interactions, many conditions can give me >15 response. I have used the response surface profiler to find the range of each interaction but this is limited to interaction between two factors at a time(?). Is there a way JMP can list all of these combinations that could give me >15 and give me a design space for all four factors?

Many thanks in Advance.

Re: How do I get a list of all combinations that can give greater than a threshold of the response in JMP DoE RSM model

Do_AuLu — Mon, 19 Feb 2024 12:21:21 GMT

Typo. Only three factors have significant main effects.

Re: How do I get a list of all combinations that can give greater than a threshold of the response in JMP DoE RSM model

SamGardner — Mon, 19 Feb 2024 15:41:00 GMT

There are a few ways to do this:

Create an output table of simulated results, randomly or systematically choosing input values and filter that output. Sometime we call that approach Monte Carlo Filtering.

Run the "Plot X filter Y" script saved to that table. Turn on the profiler simulator, set the input variables to have random uniform distribution, and then simulate to a data table.
Use the Design Space Profiler

which is described in more detail here https://community.jmp.com/t5/Mastering-JMP/Maximizing-Quality-Using-Design-Space-Profiling/ta-p/614427