topic Re: How to Save Prediction Formula for Generalized Regression Using JSL in Discussions

How to Save Prediction Formula for Generalized Regression Using JSL

ih — Thu, 02 Mar 2017 14:05:14 GMT

In partitions or linear regressions either 'Prediction Formula' or 'Save Prediction Formula' creates a new column in the data table. In the generalized regression platform neither work. In the example below, I suspect this is because I actually want to send a message to the 'Adaptive Lasso with AICc Validation' part of the Generalized Regression report instead of the top level 'Generalized Regression for Oxy'. Any idea how to do this?

dt = Open( "$SAMPLE_DATA/Fitness.JMP" )  //open sample data

//Fit model with variable reduction
lasso = dt << Fit Model(
	Y( :Oxy ),
	Effects( :Age, :Weight, :Runtime, :RunPulse, :RstPulse, :MaxPulse ),
	Personality( "Generalized Regression" ),
	Generalized Distribution( "Normal" ),
	Run(
		Fit(
			Estimation Method( Lasso( Adaptive ) ),
			Validation Method( AICc )
		)
	)
);

//Save the prediction formula to the table
lasso << Save Prediction Formula;
lasso << Prediction Formula;
lasso << Get Prediction Formula;
//None of these work, the log returns 'Fit Generalized[]' for each.

Re: How to Save Prediction Formula for Generalized Regression Using JSL

txnelson — Thu, 02 Mar 2017 16:49:36 GMT

The Save Prediction Formula needs to be part of the Fit syntax

dt = Open( "$SAMPLE_DATA/Fitness.JMP" ) ; //open sample data

//Fit model with variable reduction
lasso = dt << Fit Model(
	Y( :Oxy ),
	Effects( :Age, :Weight, :Runtime, :RunPulse, :RstPulse, :MaxPulse ),
	Personality( "Generalized Regression" ),
	Generalized Distribution( "Normal" ),
	Run(
		Fit(
			Estimation Method( Lasso( Adaptive ) ),
			Validation Method( AICc),
			Save Prediction Formula(1) )
	)
);

Re: How to Save Prediction Formula for Generalized Regression Using JSL

ih — Fri, 03 Mar 2017 13:15:27 GMT

Perfect, thank you!