2
$\begingroup$

Hello I am using SPSS trees - CART functionality for classification. I have a number of classifying variables such as Business_Size, Location, Previous_Record etc. I want to know how do I set configuration so that a particular classification variable is used at only one level of the tree. E.g say at the first level the records were classified using Location then at no subsequent level of the tree, Location should be used. Currently, at the first level sub-trees are getting formed based on location, then on the next level sub-trees are getting formed using Business_Size but on the next level again Location variable is getting used for further classification. Any help on this would be appreciated. My colleague told me that in SAS such a setting is possible. I do not have SAS. So want to know how it can be done in SPSS.

$\endgroup$
1
$\begingroup$

It is possible with SPSS AnswerTree but I believe this program isn't sold anymore. It had a million times more functionality than the current TREE command so that's kinda strange.

In SPSS Statistics there's Analyze => Classify => Tree which generates the TREE command. This does not have the possibility to "manually" build the tree. Which IMHO renders it completely worthless if it wasn't so already.

Please be aware of the highly subjective nature of such analyses. If your colleague is convinced that different variables should be forced into the tree you may end up in a long but fruitless discussion.

| cite | improve this answer | |
$\endgroup$
  • $\begingroup$ Hello, thanks for your reply. I hope I got what you are saying. I do not want to build the tree manually. I need algorithm to consider a particular variable only once. So if var A has been used to split the records at the first level then var A should not be used in any of the subsequent sub-trees. Essentially I am trying to show graphically, which variables are more significant in splitting the records. $\endgroup$ – Abhijit May 28 '14 at 8:06
  • $\begingroup$ Yes, if var_A comes up a second/third time, you manually choose a different variable for the split. This discussion kinda moved here too. $\endgroup$ – RubenGeert May 28 '14 at 8:35

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.