turn on suggestions

Auto-suggest helps you quickly narrow down your search results by suggesting possible matches as you type.

Showing results for

- JMP User Community
- :
- Discussions
- :
- Discussions
- :
- Re: Prediction Profiler Simulator Problem

Topic Options

- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content

Feb 15, 2017 3:28 PM
(1448 views)

Greetings,

I'm working with a formula in a prediction profler and using the simulator to do analysis around a selected point. For two parameters, I have found that when I update them the all the other parameters are reset to initial values that come up when I open the profiler. I was doing this via JSL scripting at first so I figured I got something wrong there. But when I manually tried to do it from the menu using the same formula, I encountered the same problem - I can update the simulator parameters manually except for the two problem ones and when I update those, all inputs reset to the values that come up when I load the prediction profiler.

Has anyone ever encountered this phenomenon? If so, how did you resolve the problem? Thanks.

Occurred as I was writing this: One thing I just noticed using a different equation with many of the same inputs is that by going outside the ranges, this triggers the reset behavior. Some of my inputs for my prediction formula are formulas themselves due to linkages of inputs. Might that be the problem? If so, is the proper resolution to simply modify my script to update the input ranges to reflect what my true ranges are?

Basically my bounds are X(0,50) -> Y. But my implementation is T(0,3) -> X(5,31) -> Y where 5 and 31 are defined due to the linkage to T. Hopefully I've made what I'm doing a little clearer.

Solved! Go to Solution.

1 ACCEPTED SOLUTION

Accepted Solutions

Highlighted
Solution

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content

Feb 23, 2017 9:51 AM
(2680 views)

Turns out the problem was that I was trying to take the profiler out of the ranges it determined from my reduced dataset.

One either needs to make sure the min and max are in the data or modify the axes ranges manually and save that into the script data.

3 REPLIES

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content

Feb 16, 2017 5:16 AM
(1427 views)

Hello,

You mention that you have inputs that refer back to formulas. One thing you can try if you haven't already is to tick the box in the lower left of the Profiler setup "Expand Intermediate Formulas" this will put things in the correct ranges. You can then go to the red hot spot by Profiler (the top one) and select Save Expanded Formulas. This will add a new column to the data table and you can then use that column to build your profiler going forward.

HTH

Bill

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content

Feb 16, 2017 10:13 AM
(1414 views)

Bill,

For one of the things I do, I do that but perhaps differently. I simply add all the prediction formulas into the profiler and that gets the desired interactions from the H(x) = F(G(x)).

It appears that I am correct that the ranges for my F(x) function (as driven by my G(x) function) are not completely capturing my space. The solution is to either: put in inputs for X for G(x) such that I get the desired F(x) ranges OR just configure the JSL to update the profiler min and max ranges to encompass what I want. I think either solution will work but I'll have to decide what works best for my application.

Thanks for the help.

Highlighted
Solution

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content

Feb 23, 2017 9:51 AM
(2681 views)

Turns out the problem was that I was trying to take the profiler out of the ranges it determined from my reduced dataset.

One either needs to make sure the min and max are in the data or modify the axes ranges manually and save that into the script data.