Deep Learning Based Hybrid Computational Intelligence Models for Options Pricing

Abstract

Options are commonly used by traders and investors for hedging their investments. They also allow the traders to execute leveraged trading opportunities. Meanwhile accurately pricing the intended option is crucial to perform such tasks. The most common technique used in options pricing is Black-Scholes (BS) formula. However, there are slight differences between the BS model output and the actual options price due to the ambiguity in defining the volatility. In this study, we developed hybrid deep learning based options pricing models to achieve better pricing compared to BS. The results indicate that the proposed models can generate more accurate prices for all option classes. Compared with BS using annualized 20 intraday returns as volatility, 94.5% improvement is achieved in option pricing in terms of mean squared error.