I'm working on a project and want to use decisions tree (because I have both catgorical and numerical values in my input and don't want to transform the categoricals variables) to predict an output varaible , but the problem is that I don't have only one output variable but 4 !
that picture is a look for my output dataset:
- The first one will can be predict with a regression decision tree .
the the rest if can be predict with a classification decision tree but with more than 2 classes so a multi-class.
So I want to know what is the best solution to approach this problem:
Using 4 differents decisions trees one tree for each variables ?
- Or using 1 multi-output decision tree?
- What are the avantages and inconvenients of each approach?