Question



Using Random Forest

Hello Dr. Ballings,

Can you explain why we need to use Random forest for the calibration assignment?
Since the data provided only contains predictions for 2 unknown models, how can
we implement the random forest process into our functions?Also, is it safe to say that
the prediction and prediction2 variables in the churn data are the score values (for
2 separate models) that we use to create bins for using the function (that we are creating)
called calibrate?

Thank you,

Shelby H.





Answers and follow-up questions





Answer or follow-up question 1

Dear Shelby,

"Can you explain why we need to use Random forest for the calibration assignment?
Since the data provided only contains predictions for 2 unknown models, how can
we implement the random forest process into our functions?"

Once you've binned your data, you will have to learn a model where the dependent is a continuous variable.
You could you OLS regression, or a decision tree, but in this case I want you to use random forest.


Also, is it safe to say that
the prediction and prediction2 variables in the churn data are the score values (for
2 separate models) that we use to create bins for using the function (that we are creating)
called calibrate?

Yes. Please use prediction or prediction2.

Michel Ballings



Sign in to be able to add an answer or mark this question as resolved.