I am running regression models (JMP15) on a dataset "by-group" - i.e. generating 100+ regression models from the same data table. Then I maximize desirability for all these groups. I would like to extract the variable values corresponding to the "maximized desirability" for each of the 100+ models (the circled values in picture below). Any suggestions on how to do that (either JSL or using the GUI are okay)?
I can do it in a tedious way (Ctrl-Prediction Profiler -> Factor Settings -> Remember Settings saves the settings for each case as "Remembered Settings", which I can then extract with Ctrl-Make Combined Data Table. However this requires me to hit "enter" 100+ times corresponding to each by-group!).
Accepted Solutions
Here is an example of doing what you want. The example creates a sample table. Then runs the Fit Model. You will have to Maximize Desirability and then run the remainder of the code to get the new output table
Names Default To Here( 1 );
dt =
// Open Data Table: Tiretread.jmp
// → Data Table( "Tiretread" )
Open( "$SAMPLE_DATA/Tiretread.jmp" );
dtSub = dt << subset( selected columns( 0 ), selected rows( 0 ), output table( "Subset" ) );
For Each Row(
:Abrasion = :Abrasion + :Abrasion*random normal(0,1);
:MODULUS = :MODULUS + :MODULUS*random normal(0,1);
:ELONG = :ELONG + :ELONG * random normal(0,1);
:HARDNESS = :HARDNESS + :HARDNESS* random normal(0,1);
);
dt << Concatenate(
dtSub,
Output Table( "Concat of Tiretread, Subset" ),
Append to first table,
Create source column
);
Close( dtSub, nosave );
// Run the Model
fm = Fit Model(
Y( :ABRASION, :MODULUS, :ELONG, :HARDNESS ),
By( :Source Table ),
Effects( :SILICA, :SILANE, :SULFUR ),
Personality( "Standard Least Squares" ),
Emphasis( "Effect Leverage" ),
Run(
:ABRASION << {Summary of Fit( 1 ), Analysis of Variance( 1 ), Parameter Estimates( 1 ),
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 )},
:MODULUS << {Summary of Fit( 1 ), Analysis of Variance( 1 ), Parameter Estimates( 1 ),
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 )},
:ELONG << {Summary of Fit( 1 ), Analysis of Variance( 1 ), Parameter Estimates( 1 ),
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 )},
:HARDNESS << {Summary of Fit( 1 ), Analysis of Variance( 1 ), Parameter Estimates( 1 ),
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 )},
Profiler(
1,
Confidence Intervals( 1 ),
Desirability Functions( 1 )
)
)
);
// Stop the JSL
stop();
// Set the Maximize Desirability and the run the code below
// Create an output table
dtMax = New Table( "Desirability",
New Column( "By Group", character ),
New Column( "Silica" ),
New Column( "Silane" ),
New Column( "Sulfer" )
);
// Populate the table getting the Desirability values from the chart
// In this example, the 2 by groups are actually just a replicate of the same
// data, so the output values will be the same.
For Each( {profile}, fm,
dtMax << add rows( 1 );
:By Group[N Rows( dtMax )] = Report( profile )[Outline Box( 1 )] << get title;
:Silica[N Rows( dtMax )] = Report( profile )["Prediction Profiler"][Number Edit Box( 1 )]
<< get;
:Silane[N Rows( dtMax )] = Report( profile )["Prediction Profiler"][Number Edit Box( 2 )]
<< get;
:Sulfer[N Rows( dtMax )] = Report( profile )["Prediction Profiler"][Number Edit Box( 3 )]
<< get;
);
Hello Jim,
Thank you very much for the code. However it gives an error in the very last section (after the Fit Least Squares is done and after it creates the empty table "Desirability") as follows:
I'm running JMP15.