Model-Free Inference of Investor Preferences: A Relative Entropy IRL Approach

TL;DR

This paper introduces a novel RE-IRL framework to infer investor preferences from market data without requiring known transition probabilities, incorporating a K-nearest neighbor method and statistical validation.

Contribution

It develops a model-free IRL approach tailored for financial markets, addressing data sparsity and validation challenges in investor preference inference.

Findings

RE-IRL effectively recovers investor reward functions from observed actions.

K-nearest neighbor approach improves behavior policy estimation.

Statistical testing framework assesses the robustness of inferred preferences.

Abstract

We present a framework using Relative Entropy Inverse Reinforcement Learning (RE-IRL) to recover investor reward functions from observed investment actions and market conditions. Unlike traditional IRL algorithms, RE-IRL is employed to account for environments where transition probabilities are unknown or inaccessible. To address the challenge of data sparsity, we utilize a $K$-nearest neighbor approach to estimate the observed behavior policy. Furthermore, we propose a statistical testing framework to evaluate the validity and robustness of the estimated results.