Machine Learning Solutions to Challenges in Finance: An Application to the Pricing of Financial Products

Abstract

The recent fast development of machine learning provides new tools to solve challenges in many areas. In finance, average options are popular financial products among corporations, institutional investors, and individual investors for risk management and investment because average options have the advantages of cheap prices and their payoffs are not very sensitive to the changes of the underlying asset prices at the maturity date, avoiding the manipulation of asset prices and option prices. The challenge is that pricing arithmetic average options requires traditional numerical methods with the drawbacks of expensive repetitive computations and non-realistic model assumptions. This paper proposes a machine-learning method to price arithmetic and geometric average options accurately and in particular quickly. The method is model-free and it is verified by empirical applications as well as numerical experiments.

Keywords: Machine learning; Finance applications; Asian options; Model-free asset pricing; Financial technology.

JEL Classification: G11, G32