Tirth, I agree that Decision Tree is not the best model for churn prediction. I used it because it is easily interpretable. With the model visualization, you can observe if the features are processed correctly. I wouldn’t use a neural network for churn prediction because it is harder to interpret.
