BookmarkSubscribeRSS Feed
question

Community Trekker

Joined:

Mar 14, 2013

Any SAS code behind the JMP Decision Trees? Help Please

Hi All,

I would like to know if JMP produces SAS Code that shows the rules behind a Decision Tree?

Many Thanks

Alice

5 REPLIES
lwx228

Regular Contributor

Joined:

Aug 14, 2018

Re: Any SAS code behind the JMP Decision Trees? Help Please

It seems that some people have asked this question, I would like to know, but it is estimated that trade secrets, will not be made public.

However, I recommend that JMP's decision tree module should be added with options: for example, do not use average as a comparison criterion.
txnelson

Super User

Joined:

Jun 22, 2012

Re: Any SAS code behind the JMP Decision Trees? Help Please

If you are asking if the Partition Platform can produce the scoring code in SAS syntax, the answer is yes.  Below is the example from the Scripting Index

     Help==>Scripting Index

Names Default To Here( 1 );
dt = Open( "$SAMPLE_DATA/Car Poll.jmp" );
obj = Partition(
	Y( :country ),
	X( :sex, :marital status, :age, :type, :size ),
	Method( "Decision Tree" ),
	Validation Portion( .2 )
);
obj << SplitBest( 2 );
code = obj << Get SAS Data Step;
Jim
lwx228

Regular Contributor

Joined:

Aug 14, 2018

Re: Any SAS code behind the JMP Decision Trees? Help Please

thank Jim!

 


Well, maybe I missed the questioner on the first floor.

I wanted to understand how the JMP decision tree worked out this result.It is not JMP that releases the resulting formula into SAS or python code.


I understand the steps of JMP release formula, and I can directly restore JMP formula with VBA full automatic and check the calculation in excel.

 

 

2018-09-06_21-50-22.png

lwx228

Regular Contributor

Joined:

Aug 14, 2018

Re: Any SAS code behind the JMP Decision Trees? Help Please

I mainly want to understand what specific steps JMP can use to calculate this result formula.
It's not just simple:obj = Partition(


JMP's decision tree does not provide many options, but I would also like to focus on whether it is feasible to judge the formula. It only USES the average value to judge the formula.

 

2018-09-06_22-14-30.png

lwx228

Regular Contributor

Joined:

Aug 14, 2018

Re: Any SAS code behind the JMP Decision Trees? Help Please

This involves algorithms, which are trade secrets.