jmp-l
[Top] [All Lists]

Scripting Help

To: jmp-l@lists.biostat.wustl.edu
Subject: Scripting Help
From: "Eric Lofgren" <eric.lofgren@gmail.com>
Date: Wed, 16 Aug 2006 11:33:01 -0400
Domainkey-signature: a=rsa-sha1; q=dns; c=nofws; s=beta; d=gmail.com; h=received:message-id:date:from:to:subject:mime-version:content-type:content-transfer-encoding:content-disposition; b=PUKirKV1QZwCYuu7Khhp7L5DperRy5KnxJQmZujajudrDsJ7IAcE2u2/ncWjZAd/eRfYooqVHxO1AHAL0XU+uyVyPh4qE9olBhTnaNPu8JKEg1nbzoHdZlFFf9l5qbG+0pHuagU8ZlewHtovelZA/pOPKdFc/ugTk4sE9GT/vuM=
First post here, hopefully it will work!

+++++

I've been attempting recently to automate what is otherwise a long and
painful task to do by hand within JMP. It amounts to grabbing values
from a Fit Model report, performing some calculations on them, and
appending them to the report itself. Right now, for the *whole* data
set, my script is working just fine. The final reporting is a bit of
brute-force code, and less elegant than I would like (I may post on
that later), but for now, it's fine.

My issue is here: The data set has a grouping variable, "Year", that I
would like to seperate the analysis by. That presents two problems:

1. How do I script a analysis to run by the "Year" variable.
Currently, the script I'm using is this:


model=Fit Model(Y( :Interest), Effects( :Name("Effect1"),
:Name("Effect2")), Personality(Generalized Linear Model), GLM
Distribution(Poisson), Link Function(Log), Overdispersion Tests and
Intervals(0), Run Model(Covariance of Estimates(1), Studentized
Deviance Residuals by Predicted(0), SendToReport(Dispatch({"Covariance
of Estimates"}, "", MatrixBox, Set Format(9, 99)))));

Which was obtained using the inbuilt mechanisms for obtaining a
script. Where would By(Year) go, and how should it be formatted? So
far, everything I have tried has really upset JMP. Should I be using
the Where command instead? If so, how?

2. How does JMP handle custom-scripted analysis for split data sets?
For example, for the whole model, I use the following script:

ests=reportmodel["Parameter ?",
       columnbox("Estimate")];
Bo=ests<<get(1);
E1Est=ests<<get(2);
E2Est=ests<<get(3);

And then, for example, use

Randominfo=Bo*E1Est*E2Est

and append that to the bottom of the GLM report, will it be retained
on *all* the reports for each year? If not, how does one go about
making sure they appear for each grouped report?



Thanks in advance for your time and assistance,
Eric
<Prev in Thread] Current Thread [Next in Thread>