Subscribe Bookmark RSS Feed

JMP Sript for Robust Regression with SAS

Hi. I'm trying to automate a multivariable robust regression (many X's one Y) through the JMP add-in. So far, what I have sort of works, but it's not quite as slick as I'd like it to be. Here's what I've been using:

//===========================
// Run Robust Regression
//===========================

SASJMP$checkDataTableOpen();
robustReg$jmpDataTable = Current Data Table();

//===========================
// Create the dialog window
//===========================
robustReg$app();

//===========================
// use preselected roles from data table
//=================
robustReg$applyPreselectedRoles();

// Set defaults
//=============
robustReg$setModelDefaults();
robustReg$setMethodDefaults();
robustReg$setTablesDefaults();

If( robustReg$validate(), robustReg$submit());

Can anyone tell me how to change the defaults for Model, Method, and Tables? I've been using the wdw() command to get and close windows as they pop up, which is less than efficient.

Also, can anyone recommend a way to keep this for loop from eating up all my memory? I've been running this script inside a Local({variables}, body) function but it steadily increases my memory until my whole PC hangs. The task manager says JMP is only running ~200MB but my PC ends up running 7.2 GB total. Is there a "garbage collection" feature I need to set in SAS java runtime?

Oh, and one more item - what's the best way to clean up the SAS Submit folder that contains all of the SAS results? I could create a function reference to kernel32.dll and delete them that way, but seems less than practical.

That's quite a few questions. Any ideas would help as I am a JSL novice but am quickly beginning to see the light!

- shwaZmoto