Adding three models that are a simple wrapper around existing LGBM, XGB and GLM models, using transformed outcome to predict uplift for binary classification uplift models.
In addition, this adds also the AUUC (area under uplift curve) scorer. Some code in the scorer is copied over from https://github.com/uber/causalml and is properly acknowledged. As that code is also Apache2, I hope it's okay.
Adding three models that are a simple wrapper around existing LGBM, XGB and GLM models, using transformed outcome to predict uplift for binary classification uplift models.
In addition, this adds also the AUUC (area under uplift curve) scorer. Some code in the scorer is copied over from https://github.com/uber/causalml and is properly acknowledged. As that code is also Apache2, I hope it's okay.
An example notebook how to use the recipes: https://github.com/h2oai/dai-domain-solution-recipes/blob/master/uplift/criteo-minimal.ipynb