Building models with SAS Enterprise Miner, SAS Factory Miner, SAS Visual Data Mining and Machine Learning or just with programming

Decision tree applied with ternary partition of continuous variables

Reply
Occasional Contributor
Posts: 12

Decision tree applied with ternary partition of continuous variables

I have a modelling excercise where I want to apply a decision tree model to a dataset with continuous variables. The target is a binary outcome. At each node I want division into 3 cases, pure 1; mixed 1,0; and pure 0. and the next split applied to the mixed case, in the same way. The resulting tree looks like a main trunk, with one leaf on either side of each node. Is it possible to constrain the decision tree process to achive this?
SAS Employee
Posts: 35

Re: Decision tree applied with ternary partition of continuous variables

The software does not support that constraint. In fact, there may be few data sets that contain candidate splits that satisfy that constraint. You may want to use the interactive tree function to explore splits and find out if that is even possible.
Occasional Contributor
Posts: 12

Re: Decision tree applied with ternary partition of continuous variables

David, Thanks, I'll have a play. I've also started coding an algorithm to do it directly.
Ask a Question
Discussion stats
  • 2 replies
  • 174 views
  • 0 likes
  • 2 in conversation