Option Pricing With Machine Learning

49 Pages Posted: 25 Nov 2019 Last revised: 5 Dec 2019

See all articles by Daniel Alexandre Bloch

Daniel Alexandre Bloch

Université Paris VI Pierre et Marie Curie

Date Written: November 12, 2019

Abstract

An option pricing model is tied to its ability of capturing the dynamics of the underlying spot price process. Its misspecification will lead to pricing and hedging errors. Parametric pricing formula depends on the particular form of the dynamics of the underlying asset. For tractability reasons, some assumptions are made which are not consistent with the multifractal properties of market returns. On the other hand, non-parametric models such as neural networks use market data to estimate the implicit stochastic process driving the spot price and its relationship with contingent claims.

When pricing multidimensional contingent claims, or even vanilla options with complex models, one must rely on numerical methods such as partial differential equations, numerical integration methods such as Fourier methods, or Monte Carlo simulations. Further, when calibrating financial models on market prices, a large number of model prices must be generated to fit the model parameters. Thus, one requires highly efficient computation methods which are fast and accurate.

Neural networks with multiple hidden layers are universal interpolators with the ability of representing any smooth multidimentional function. As such, supervised learning is concerned with solving function estimation problems. The networks are decomposed into two separate phases, a training phase where the model is optimised off-line, and a testing phase where the model approximates the solution on-line. As a result, these methods can be used in finance in a fast and robust way for pricing exotic options as well as calibrating option prices in view of interpolating/extrapolating the volatility surface. They can also be used in risk management to fit options prices at the portfolio level in view of performing some credit risk analysis.

We review some of the existing methods using neural networks for pricing market and model prices, present calibration, and introduce exotic option pricing. We discuss the feasibility of these methods, highlight problems, and propose alternative solutions.

Keywords: Machine Learning, Supervised Learning, Option Pricing, Calibration

Suggested Citation

Bloch, Daniel Alexandre, Option Pricing With Machine Learning (November 12, 2019). Available at SSRN: https://ssrn.com/abstract=3486224 or http://dx.doi.org/10.2139/ssrn.3486224

Daniel Alexandre Bloch (Contact Author)

Université Paris VI Pierre et Marie Curie ( email )

175 Rue du Chevaleret
Paris, 75013
France

Here is the Coronavirus
related research on SSRN

Paper statistics

Downloads
185
Abstract Views
564
rank
171,139
PlumX Metrics