UTBoost: Gradient Boosted Decision Trees for Uplift Modeling
This work addresses the problem of predicting causal effects for managers in marketing or policy-making, though it appears incremental as it builds upon existing GBDT techniques.
The paper tackles the challenge of accurately estimating the incremental impact of actions on customer outcomes in uplift modeling by introducing two novel modifications to Gradient Boosted Decision Trees (GBDT) that sequentially learn causal effects, achieving substantial improvements over baseline models in large-scale experiments.
Uplift modeling comprises a collection of machine learning techniques designed for managers to predict the incremental impact of specific actions on customer outcomes. However, accurately estimating this incremental impact poses significant challenges due to the necessity of determining the difference between two mutually exclusive outcomes for each individual. In our study, we introduce two novel modifications to the established Gradient Boosting Decision Trees (GBDT) technique. These modifications sequentially learn the causal effect, addressing the counterfactual dilemma. Each modification innovates upon the existing technique in terms of the ensemble learning method and the learning objective, respectively. Experiments with large-scale datasets validate the effectiveness of our methods, consistently achieving substantial improvements over baseline models.