Optimal Stopping with Gaussian Processes

TL;DR

This paper introduces Gaussian Process-based algorithms for fast approximate optimal stopping in financial time series, leveraging structural properties like mean reversion to analytically evaluate value functions and outperform benchmarks.

Contribution

The paper presents a novel family of Gaussian Process algorithms that enable analytical evaluation of optimal stopping policies in financial data, improving speed and accuracy over existing methods.

Findings

Outperforms state-of-the-art deep learning benchmarks

Effective on multiple financial datasets including stock prices and yield rates

Quantifies uncertainty in value functions through model propagation

Abstract

We propose a novel group of Gaussian Process based algorithms for fast approximate optimal stopping of time series with specific applications to financial markets. We show that structural properties commonly exhibited by financial time series (e.g., the tendency to mean-revert) allow the use of Gaussian and Deep Gaussian Process models that further enable us to analytically evaluate optimal stopping value functions and policies. We additionally quantify uncertainty in the value function by propagating the price model through the optimal stopping analysis. We compare and contrast our proposed methods against a sampling-based method, as well as a deep learning based benchmark that is currently considered the state-of-the-art in the literature. We show that our family of algorithms outperforms benchmarks on three historical time series datasets that include intra-day and end-of-day equity stock prices as well as the daily US treasury yield curve rates.