I have built a random forest regression model for credit underwriting. But the business doesn't appreciate a black-box approach. So, using treeinterpreter, I have found the path taken for a prediction. But is there a physical interpretation of the contribution values. I understand that these values mean that, taking that path, the output changed by some value(positive/negative). But is there any intuitive explanation to these values that is feature-specific? Something that will make our decision tilt towards features with positive contribution values.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.