cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 
Browse apps to extend the software in the new JMP Marketplace
Choose Language Hide Translation Bar
ECGA94
Level I

Simulate A Definitive Screening Design 1000 times.

Hello,

 

I am trying to fit a definitive screening design with a dataset that I already have. I would like to write a JSL script that would fit the design 1000 times. Would anyone be able to tell me how to go about doing this?

 

I have a Y variable of interest, and X1, X2, X3, and X4 as my four x variables. 

 

Thanks in advance for your help!

4 REPLIES 4
P_Bartell
Level VIII

Re: Simulate A Definitive Screening Design 1000 times.

Some more context and information is needed. You say you already have a 'data set'. Is the data set in the format/design of a DSD? Why would you want to fit a model 1000 times? If the data set already exists you'll just get the same analysis for all 1000 iterations. Are you planning to use the Fit DSD platform or some other platform like Fit Model? And maybe some explanation of the practical problem at hand? And how you actually acquired the data set?

 

It almost sounds like what you've got is model already...and now you'd like to simulate that model 1000 times. Best place in JMP to accomplish this is in the Profiler. You don't need JSL to do that...just specify the number of simulations within the workflow of the simulator and you'll get a distribution of Y's.

 

Or am I all just confused?

ECGA94
Level I

Re: Simulate A Definitive Screening Design 1000 times.

Hi P_Bartell,

 

Thank you for your response. For more context, what we are trying to do is simulate new data for our outcome Y variable 1000 times, and we are looking to fit the DSD model for each simulation.

 

- We have simulated on a theoretical case, which is the equation below:

 

Y = 1*X3 + 1*X1*X1 -2*X4*X4 + 1*X1*X2 + Random Normal (0,1)

 

Here is a screenshot of our DoE design, which is the default design for 4 factors with center points and 4 extra runs:

ECGA94_0-1713961153112.png

After the simulation is run, we want to count how many times the correct factors were retained in the models. We have been able to do this using the Stepwise platform, but for completeness, we would also like to be able to do it on the default Fit Definitive Screening Design platform.

 

We are doing this for educational purposes.

Re: Simulate A Definitive Screening Design 1000 times.

Please see the documentation for the built-in simulation feature in JMP Pro.

ECGA94
Level I

Re: Simulate A Definitive Screening Design 1000 times.

Hi Mark,

 

I don't have access to JMP Pro, only JMP 17, so we're trying to find a way to achieve this using JSL scripting. However, there isn't much information in the scripting index for the default Fit Definitive Screening Design platform.