Fast Learning of Optimal Policy Trees

TL;DR

This paper introduces a faster algorithm for learning optimal policy trees using discrete optimization, significantly reducing runtime and providing an accessible R package for practical application.

Contribution

It develops an improved version of the policytree method with enhanced computational efficiency and releases an R package for public use.

Findings

Runtime improved by nearly 50 times

Effective in finite sample settings

Open-source R package available

Abstract

We develop and implement a version of the popular "policytree" method (Athey and Wager, 2021) using discrete optimisation techniques. We test the performance of our algorithm in finite samples and find an improvement in the runtime of optimal policy tree learning by a factor of nearly 50 compared to the original version. We provide an R package, "fastpolicytree", for public use.