Fit Model: Script vs Platform Process Speed

Question

I am having an issue with the Fit Model platform (JMP 14.3). When I run the Fit Model through the platform window for a specific analysis, it will take about 10 seconds to complete. However, if I copy the script and re-run it on the same table, it will take nearly 10 minutes to complete. It takes even longer if I convert the script to a string and use the Eval(Parse()) method. Has anybody else experienced this?

anthonee_g · Answer

I am unclear on the JMP jargon, sorry. To answer your first question, I am trying to use the Standard Least Squares personality. I'm not sure if it changes things, but it's defined as "The Standard Least Squares personality within the Fit Model platform".

Here is the script that I am using:

Function = Function(
	{dt = Current Data Table(), by = {:Name( "FileName" ), :Name( "Wafer#" ), :Name( "Shot#" )}, y = {:X_Mesr,
	:Y_Mesr}, effects = {:Name( "X" ), :Name( "Y" )}, where = ":Filter == \!"Filter\!""},

a = "Data Table( " || Char( dt ) || " ) << Fit Model(
		By( Eval( " || Char( by ) ||
	" ) ),
		Y( Eval( " || Char( y ) || " ) ),
		Effects( Eval( " || Char( effects ) ||
	" ) ),
		Center Polynomials( 0 ),
		Personality( \!"Standard Least Squares\!" ),
		Emphasis( \!"Effect Leverage\!" ), 
		Where( Eval("
	 || Char( where ) || " ) ), ";

b = "";

For( i = 1, i <= N Items( y ), i++,

b1 = Char(
			"" || Char( y[i] ) ||
			" << {Summary of Fit( 1 ), Analysis of Variance( 1 ), Parameter Estimates( 1 ), Lack of Fit( 0 ), Scaled Estimates( 0 ),
				Plot Actual by Predicted( 1 ), Plot Regression( 0 ), Plot Residual by Predicted( 1 ), Plot Studentized Residuals( 0 ),
				Plot Effect Leverage( 1 ), Plot Residual by Normal Quantiles( 0 ), Box Cox Y Transformation( 0 ), Prediction Formula( 1 ),
				Residuals( 1 )},"
		);

b = b || b1;
	);

c = "Run( " || b || " )";

d = a || c;

Show( d );

Eval( Parse( d ) );

)

Mark_Bailey · Answer

First of all, Fit Model is not a platform. It is a launch dialog for many platforms such as Fit Least Squares or Nominal Logistic Regression. Which platform are you actually using?

Second, can you share the script for this launch? It is difficult to determine the root cause of your problem as there are many things that can affect the run time.

I have not experienced anything like this difference in run time but someone else might have.