I want to build an Uplift Model for multiple treatments. To get a good model, I would like to use boosting. How is it possible to use boosting with uplift modeling although we can't really calculate residuals and build our next tree on the last one (Causal inference)? It seems to be impossible to improve our current tree if we don't have a "true" uplift value for each observation, which we can use to calculate residuals.


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.