SUPR
Machine Learning Approaches for High-Dimensional Option Pricing and Hedging
Dnr:

NAISS 2025/22-690

Type:

NAISS Small Compute

Principal Investigator:

Ying Ni

Affiliation:

Mälardalens universitet

Start Date:

2025-05-07

End Date:

2026-06-01

Primary Classification:

10105: Computational Mathematics

Webpage:

Allocation

Abstract

This project is in the field of financial mathematics and in particular computational finance. The aim is to advance the fields of option pricing and hedging by applying state-of-the-art machine learning techniques, with a focus on computationally intensive financial derivatives. The first component addresses the pricing of Bermudan basket options in high-dimensional settings. The second component of the project centers on option hedging. Building on the deep hedging paradigm, we test a novel hedging framework that integrates machine learning tools with user-centered explainability and high hedging performance under advanced stochastic market models. Both components of the project involve extensive simulations, high-dimensional modeling, and training of large-scale machine learning models, necessitating access to high-performance computing resources. The proposed research has the potential to significantly advance the computational methods used in quantitative finance, particularly in scenarios where dimensionality and model complexity pose major challenges.