Record wise/Customer wise variable importance using Tree Ensemble Learner


I build a model to predict Churn. I also created a report using BIRT Reporting. I used Tree Ensemble learners to get variable importance. I like to see the variable importance of each Customer. Below is the screenshot of the table view I want to add the essential variables by each Customer. Is this possible?
knime report image.docx (2.6 MB)

Hi there @shiuleec,

Welcome to KNIME Community Forum!

to my understanding that is not possible. What you could do is to segment your customers first and then for each group of customers build a churn model. This way you can determine variable importance per each group.


This topic was automatically closed 182 days after the last reply. New replies are no longer allowed.