I have built a propensity model, which gives out probabilities of a customer paying given a collection intervention using a xgboost model. The model has an AOC-ROC of 81% with an accuracy of 77% having TPR of 88%. I deployed the model to sort the calling list by the probabilities of a customer paying given an intervention. When I do a pre vs post analysis, there's hardly any impact of the model. Can anyone tell what can I be missing? Is there a problem with the model or something else is hindering the impact?


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.