Simulation Experiment improvements

[shampton82]

I'd like to suggest the following to updates to simulation experiment in the Simulator

• Have the order of the prediction profiler factors in the Gaussian Process Model match the original prediction profiler factor order
• For JMP Pro users, have the fast gasp option be turned on by default, you have to modify the saved script in the Gaussian data table currently.

Thanks!

Steve

[Sarah-Sylvestre]

@shampton82 - Thank you for your suggestion! We have captured your request and will take it under consideration.

[SamGardner]

@shampton82 I don't understand what you mean by "Have the order of the prediction profiler factors in the Gaussian Process Model match the original prediction profiler factor order".  Can you explain better what you experience when you use the Profiler Simulatar in Gaussian Process Model?

Fast GASP is a checkbox option in JMP Pro version 16 in the launch dialog, but there is not a preference to have to turned on by default. Also, if there are more than 2500 rows in the data table, Fast GASP is selected automatically.  Is that what you are asking for, to have a preference to always have Fast GASP selected?  

[shampton82]

Hey Sam,

Thanks for reaching out!

For question 1, here you can see i reordered the main analysis profiler but when I run the simulation experiment(SE) the profiler is in the normal ordering.  In this case it's not a big deal as there are only two factors but when there are a bunch this makes it get harder to switch back and forth between the platforms.

For question 2, I have to type this in each time I run the simulation experiment(SE) as fast gasp wont default on for the standard 128 runs by default.

Ideally if you have JMP Pro Fast Gasp could always be on by default in the SE.

[SamGardner]

@shampton82 please contact JMP Technical Support regarding the Fast Gasp issue, if you have not already, so we can investigate this. I am not seeing the same behavior that you describe. You can mention this post and that will allow us to keep the tech support case linked to it.  

There is a semi-complicated workaround for the factor ordering is to capture the script for the first profiler that you re-ordered, and copy the "Reorder X Variables()" message, and then use that in subsequent profiler scripts.   We are considering ways we can improve the user interface and user experience for the profilers, and we will add this to the list of issues we may try to address.  

[shampton82]

Thanks for the update Sam!

I opened this ticket per your suggestion TS-00026468.

[SamGardner]

@shampton82 thank you for the follow-up with technical support.  We will look into the idea of enabling Fast GASP by default in the script that is saved to the data table in a Simulation Experiment that you can create using the Profiler simulator.  

[SamGardner]

We are looking to provide this in the next major release of JMP.  

[shampton82]

@SamGardner Nice!!