Decision tree (C5.0 algorithm): limit tree depth

[dsokhi]

Hi,

I am having a question about decision tree tool in Alteryx. Below is the context of the problem I am looking to solve.

We have a questionnaire of 14 questions (predictors) and a dependent variable. All the predictors are yes/no questions except one which is a ordinal variable. The dependent variable is ordinal variable (empirically).

I want to create a decision tree where I can get to the leaf node by asking at the most six out of 14 questions. I am aware that this can be done using rpart algorithm (using option of maximum allowed depth), however I am more interested in C5.0 algorithm as it performs better on my test dataset.

Can you please help or direct me to the relevant resources which can explain how I can achieve this. Many thanks!

Kind regards,

Dilbag

[OldDogNewTricks · Accepted solution]

Add the 'Decision Tree' tool to the canvas.  Select your target variable.  Click the 'Customize' button toward the middle of the setup page.  Change the drop down from 'rpart' to 'C5.0'.  On this page there are other options you can tune.