Decision Tree (Attribute result)

Mac2020

Helllo,
I have 10 attributes input into the model of decision tree but when I ran decision tree model, result automatically deleted some attributes.
If I want to remain all 10 attributes in the result in decision tree model.
How can I do?
Thanks and rgds, 

BalazsBaranyRM

Hi @Mac2020,

selecting relevant attributes is the main point of the decision tree algorithm. 

You might switch off all pruning options in the decision tree. That will lead to a much more complex tree, which is probably overfitted, but it could contain more attributes. There is no guarantee to keep all attributes.

If you're interested in the importance of all attributes in multiple tree models, you could try the operator Weight by Tree Importance. You have to build a Random Forest model (which consists of many decision trees, built from randomly selected attributes) and put the result into Weight by Tree Importance. This will give you a better idea on the attributes' helpfulness in the model.

Regards,
Balázs

BalazsBaranyRM

Hi @Mac2020,

selecting relevant attributes is the main point of the decision tree algorithm. 

You might switch off all pruning options in the decision tree. That will lead to a much more complex tree, which is probably overfitted, but it could contain more attributes. There is no guarantee to keep all attributes.

If you're interested in the importance of all attributes in multiple tree models, you could try the operator Weight by Tree Importance. You have to build a Random Forest model (which consists of many decision trees, built from randomly selected attributes) and put the result into Weight by Tree Importance. This will give you a better idea on the attributes' helpfulness in the model.

Regards,
Balázs

Mac2020

Hi BalazsBarany
Thank you  for your suggestion.