Abstract: Option pricing is of key importance for stock markets and traders to reduce risk, avoid loss and on the other hand speculate on stock price movements. This work explores the efficacy of using artificial neural network approach in call option pricing. We built a multilayer perceptron model trained it with real market option contracts data and tested it in option data originated from fifty S&P 500 stocks. In our approach both training and testing data are market oriented and this is a unique contribution to existing research, where training is usually based on artificially generated data. Our findings demonstrate that multilayer perceptron performs very well in actual market data and is competitive to Black-Scholes pricing formula. Further exploration and experimentation is required, however, so machine learning approaches reach required robustness and become less ad hoc and data sensitive. Despite its limitations, it is a very promising approach and can play a substantial role in option pricing, provided that it is supported by relevant software solutions.
Keywords: stock option pricing, artificial neural network, multilayer perceptron.
Title: Multilayer Perceptron model efficacy for S&P 500 Stock Option Pricing
Author: Georgios Rigopoulos
International Journal of Computer Science and Information Technology Research
ISSN 2348-1196 (print), ISSN 2348-120X (online)
Vol. 11, Issue 3, July 2023 - September 2023
Page No: 153-159
Research Publish Journals
Website: www.researchpublish.com
Published Date: 12-September-2023
