topic Re: Fit Y by X function: return equation to JMP sheet in Discussions

Fit Y by X function: return equation to JMP sheet

ndowidar — Tue, 18 Jan 2022 21:27:30 GMT

Hello,

I'd like to use my data to identify the concentration at say 1x that would be the optimal concentration. So I'm trying to use the JMP Fit Y by X feature to find the equation and how it relates to the two variables I have. The Fit Y by X feature works great but now I'd like to return the equation and values to the data table. Since the data I have returns these values 100s of times, is there an easy way of returning these to the data table so that I can actually use the equations to calculate the optimal concentrations I'm after?

Thanks so much for your help!

Using JMP 16

Re: Fit Y by X function: return equation to JMP sheet

Mark_Zwald — Tue, 18 Jan 2022 22:07:04 GMT

You can save the prediction formula of any data model you fit in a JMP report. To do with Fit Y by X, you would first run the platform with your X and Y variables, then fit the model you want to fit (ex Fit Line), then under the graph you will see a red triangle next to the legend of the Fitted Models. Click that triangle and choose "Save Predicteds". That will create a formula column in your data table of that model.
Here is an example script of that:

dt = Open( "$SAMPLE_DATA/Big Class.jmp" );
obj = dt << Bivariate( Y( :Weight ), X( :Height ) );
obj << Fit Polynomial( 1, {Save Predicteds} );

Re: Fit Y by X function: return equation to JMP sheet

Jeff_Perkinson — Tue, 18 Jan 2022 22:08:42 GMT

The Save Predicteds command from the red triangle menu will save a Formula column to the data table.

If you're looking for optimization, don't overlook the Profiler. It is tailor-made for that and will work with the formula you save from Fit Y by X.

Re: Fit Y by X function: return equation to JMP sheet

ndowidar — Wed, 19 Jan 2022 04:33:03 GMT

Thank you so much for the help! I see now that I can save the equation as a formula. I can go down through the graphs one at a time and save the predictions and that works great. I realize though that I should probably learn how to write a script to do that instead! I'd have to do this a few hundred times.... so thank you so much for that script example. Maybe I can use it by swapping the variables that are in my table...? I'll definitely give a shot.

I've never written a script before. Would I just open a new script and try it out or is there somewhere in the column that I can add the script?

If you have any other suggestions, I'd appreciate the guidance.

Thanks again!

Re: Fit Y by X function: return equation to JMP sheet

ndowidar — Wed, 19 Jan 2022 04:37:17 GMT

Thanks so much for the help. I'll definitely look into that Profiler option as well. What I have to do will have to be done hundreds of times, so maybe the Profiler will have a feature to just apply the equation from the data table.

I feel like I'm getting so close to figuring this out, it's very exciting to learn more about JMP!

Thanks again!

Re: Fit Y by X function: return equation to JMP sheet

ndowidar — Wed, 19 Jan 2022 15:44:33 GMT

I found a way to add the script as a formula for a column. I followed these steps but it's returning an error.

1.If I do that, I think I have to remove the first line of the script since it's a formula in the the window itself; i.e there's no need to write a function to open the data set, correct?

Now all I'd have in the formula box are the two lines of code:

obj = dt << Bivariate( Y( :Weight ), X( :Height ) );
obj << Fit Polynomial( 1, {Save Predicteds} );

2.Then all I should have to do is swap the Y and X variables with the ones in my column.

3. Is the Fit Polynomial equivalen to the Fit Line option or should I re-write that at "Fit Line".

Result: After I did that I got an error: Send expects scriptable object.

Any suggestions?