Optimal Policy Trees
This work addresses the problem of generating interpretable and effective prescription policies for decision-makers in various domains, offering a robust solution for data-driven recommendations.
This paper introduces Optimal Policy Trees, a method that learns interpretable, tree-based prescription policies directly from data by integrating counterfactual estimation with globally-optimal decision tree training. The method is highly scalable, handles both discrete and continuous treatments, and demonstrates best-in-class performance across various synthetic and real-world datasets.
We propose an approach for learning optimal tree-based prescription policies directly from data, combining methods for counterfactual estimation from the causal inference literature with recent advances in training globally-optimal decision trees. The resulting method, Optimal Policy Trees, yields interpretable prescription policies, is highly scalable, and handles both discrete and continuous treatments. We conduct extensive experiments on both synthetic and real-world datasets and demonstrate that these trees offer best-in-class performance across a wide variety of problems.