So I'm trying to build an ID3 decision tree but in sklearn's documentation, the algo they use is CART. But I also read that ID3 uses Entropy and Information Gain to construct a decision tree.
Tried dtree=DecisionTreeClassifier(criterion='entropy') but the resulting tree is unreliable.
What parameters should I edit in DecisionTreeClassifier() to get an ID3 decision tree?
Dear Leejames Merlas,
ID3 constructs multiway trees (on categorical features) and CART constructs binary trees, and scikit-learn implements CART as you mentioned.
If you want ID3 as a result of a need to use categorical features with more than two possible values, you may convert these categorical features to dummy features (feature is splitted into binary features as many as the possible values, only one of the features is 1 in any case).
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Programming Languages
- Python