Replication package for: Machine Learning for Dynamic Incentive Problems

This package contains replication files and source code for the computational results in "Machine Learning for Dynamic Incentive Problems," by Philipp Renner and Simon Scheidegger, to be published in the Review of Economic Studies; the README file discusses how to utilize the provided Python code, which implements Gaussian process regression and Bayesian active learning to solve the dynamic adverse selection models featuring up to ten types and limited overreporting.