Machine Learning for High-Dimensional Dynamic Stochastic Economies

Abstract

We present the first computational framework that can compute global solutions to very-high-dimensional dynamic stochastic economic models on arbitrary state space geometries. This framework can also resolve value and policy functions' local features and perform uncertainty quantification, in a self-consistent manner. We achieve this by combining Gaussian process machine learning with the active subspace method; we then embed this into a massively parallelized discrete-time dynamic programming algorithm. To demonstrate the broad applicability of our method, we compute solutions to stochastic optimal growth models of up to 500 continuous dimensions. We also show that our framework can address parameter uncertainty and can provide predictive confidence intervals for policies that correspond to the epistemic uncertainty induced by limited data. Finally, we propose an algorithm, based on this framework, that is capable of learning irregularly shaped ergodic sets as well as performing dynamic programming on them.