Developing an optimized artificial intelligence model for S&P 500 option pricing: A hybrid GARCH model

摘要

In this paper, we propose two hybrid models to release some limitations and enhancement of the results. In this regard, three popular GARCH-typc models are utilized for more accurate estimating of volatility, as the most important parameter for option pricing. Furthermore, the two non-parametric models based on Artificial Neural Networks and Neuro-Fuzzy Networks tuned by Particle Swarm Optimization algorithm are proposed to price call options for the S&P 500 index. By comparing the results obtained using these models, we conclude that both Neural Network and Neuro-Fuzzy Network models outperform the Black-Scholes model.