Save script for Fit Model with by group

[#1]

Hi JMPers,

I'm having a problem when saving a script to create multiple models that have been generated by Fit Model using a by-group. What I do is basically the following:

Use Fit Model:

In the second step I will refine the different models for each group:

Now I have two models. I would then like to save a script that reproduces this report. Sadly when saving the script using the typical way (Hotspot => Save Script (All Objects) it gives me the following output:

This is problematic because now I cannot righclick into the tables and "Save as Combined Data Table". I'm then missing the "Group"-Information. 

Would be very happy to learn how to do this. 

Best regards,

Sebastian

[#2]

I just found out I can split my table so that each group is a represented by an individual response column and then it works. 

Would still like to know if it is possible to do it the other way around.

[#3]

Not sure when this was introduced (which version) but I believe the By-Group script will do what you need:

[#4]

What JMP does with "To Data Table (All Objects)" when you have a By group is to basically create a separate report for each level of the By group and concatenate them all together within a vertical list box.  ih is right that at least with JMP 13 there is a "Save-By-Group Script" option. 

If you're not on 13 yet, then here's how I used to do it.  Save a script to a data table the regular way (not for all objects, you'll just have to do more modifications to the script if you do).  Then, edit the script to add a Group() statement within the Fit Model() statement.

For example, if saving the script to data table gets you something like this:

Fit Model(
	Y( :Y ),
	Effects( :X ),
	Personality( "Standard Least Squares" ),
	Emphasis( "Effect Screening" ),
	Run(
		Profiler( 0 ),
		:Y << {Summary of Fit( 1 ), Analysis of Variance( 1 ),
		Parameter Estimates( 1 ), Effect Details( 0 ), Lack of Fit( 0 ),
		Sorted Estimates( 0 ), Scaled Estimates( 0 ), Plot Actual by Predicted( 1 ),
		Plot Residual by Predicted( 1 ), Plot Residual by Row( 1 ),
		Plot Studentized Residuals( 0 ), Plot Effect Leverage( 0 ),
		Box Cox Y Transformation( 0 ), Show VIF( 1 )}
	),
	Where( :Group == "A" )
)

Delete the Where() clause and add a By() clause using the same column up by the other column specifications like so:

Fit Model(
	Y( :Y ),
	Effects( :X ),
	By( :Group ),
	Personality( "Standard Least Squares" ),
	Emphasis( "Effect Screening" ),
	Run(
		Profiler( 0 ),
		:Y << {Summary of Fit( 1 ), Analysis of Variance( 1 ),
		Parameter Estimates( 1 ), Effect Details( 0 ), Lack of Fit( 0 ),
		Sorted Estimates( 0 ), Scaled Estimates( 0 ), Plot Actual by Predicted( 1 ),
		Plot Residual by Predicted( 1 ), Plot Residual by Row( 1 ),
		Plot Studentized Residuals( 0 ), Plot Effect Leverage( 0 ),
		Box Cox Y Transformation( 0 ), Show VIF( 1 )}
	)
)

 You should now reproduce the Fit Model report broken out by each level of By() as desired.

[#5]

Thanks for the tips. I am on JMP 13.2.1 Sadly the "Save-By-Group-Script"-option disappears in the moment where I use the Effect Summary to remove individual effects from one model.

Kind of the same problem for the scripting approach. I will have a lot of individual fit-model statements and I don't know how to put them together. Especially as we are talking of more than 20 groups which would be a lot of work that I would like to avoid.

Best,

Sebastian

[#6]

Not sure exactly what workflow you envisage, but the thread implies a 'semi-automated' approach in which (someone!) refines each of the models 'by hand'.

But regarding: "This is problematic because now I cannot righclick into the tables and "Save as Combined Data Table". I'm then missing the "Group"-Information." (and as you know already) it would be easy enough to reinstate this column programaically.

Using 'Big Class', I made:

 

so one could take the 'Intercept' cell as the indication that this and subsequent rows are for a new level of the 'By' group.