How to get equation from fit model

[#1]

When fitting a model in JMP, it must create an equation to predict the behavior of the response variable from the parameters.  Where can that equation be found?

Thanks,
Matt

[#2 · Accepted solution]

Which  platform are you using?  If you are using Fit Model, under the red triangle you will find options to display the formula, or to save the prediction formula to the data table.  This is for the standard least squres.  Different personalities will have options for doing similar things.

Under Fit Y by X there is an option under the red triangles to save predicteds.

The Partitioning platform allows for the "Save Predicteds".  

If you can provide which modeling platform you are using, and you can not find the red triangle option for the predicted formula, repost to the discussion group with the specifics and I am sure the answers will be provided.

[#3]

Thanks very much, Jim.  I'm using Fit Model and Show Prediction Expression appears to be what I'm looking for.  Is it possible to copy this out as text?  I have a number of expressions I need to implement in Python and they're all fairly complex, but none of the options seem to allow copy/paste.

Matt

[#4]

Matt,

The "Show Prediction Expression" displays an image(Picture Box()), not a text field.  However, it is a simple step, to generate the Prediction Formula and to capture it.  See the script below

Names Default To Here( 1 );
dt = Open( "$SAMPLE_DATA\Big Class.jmp" );

fm = dt << Fit Model(
	Y( :height ),
	Effects( :weight ),
	Personality( "Standard Least Squares" ),
	Emphasis( "Effect Leverage" ),
	Run(
		:height << {Summary of Fit( 1 ), Analysis of Variance( 1 ), Parameter Estimates( 1 ),
		Plot Actual by Predicted( 1 ), Plot Residual by Predicted( 1 ),
		Plot Effect Leverage( 1 )}
	)
);

fm << prediction formula;

equation = Char( Column( dt, N Cols( dt ) ) << get formula );

Show( equation );

[#5]

Hi Jim,

Unfortunately, my experience with JMP is entirely interactive (not programmatic), and I couldn't make this work.  I tried replacing "$SAMPLE_DATA\Big Class.jmp" with my file, and everything inside Fit Model () with the script for my model, but that wasn't enough.  There really isn't any way to get the expression(s) as text within the JMP GUI without writing a separate script?

Thanks,
Matt

[#6]

You can easily do this interactively.

After you have run your Fit Model, go to the Red Triangle and select

     Save Columns==?Prediction Formula

It will create a new column in your data table.

Go to the new data column

Go to the column header for the new column, and right click on it

Select

     Formula

The formula editor will popup.

Double click on the blank space within the blue outline around the formula

It will popup a window you can copy from

[#7 · Accepted solution]

Ah, excellent, thanks again.  Using the formula from the new column works well.

[#8]

I'm working on something similar. I needed to get the published python code so I could save it as a .py and use it later.

dt = Open( "/Users/bywing/irisr.csv",End Of Field( Tab, CSV( 1 ) ) );//make sure preferences open CSV as standard CSV
model = dt << Fit Model(
	Y( :Name ),
	Effects( :Sepal length, :Sepal width, :Petal length, :Petal width ),
	Personality( Nominal Logistic ),
	Run
);
fd = model << Publish Probability Formulas;

fd << Generate Python Code;
Wait( 0 );
cw = Window( "Python Code Output" );
ed = cw[Script Box( 1 )];
code = ed << get text;
       
Save Text File( "irismodel.py", code ); // saves generated python code as python script);
Window( "Report: Formula Depot" ) << close window;

Note, this script is working with a CSV version of the Iris data, not a JMP table. not something operational, just a short POC

[#9]

m100psi: If you are running JMP Pro, you can publish your prediction equation to the Formula Depot and then from the Formula Depot hot spot, choose the "Generate Python Code" option to write the scoring code needed for Python.

[#10]

Thanks, Peter, I will try this next.  I'm running JMP Pro 11.1.1.

Matt