Use of neural networks in forecasting financial market

This paper studies solutions for forecasting option prices in a volatile financial market. It reviews a mathematical model based on traditional Black-Scholes parametric solution. Then, uses neural networks and compares the results with the conventional method. Twenty year data from S&P 500 index call option prices was used in this study. Initially simple neural network was implemented. The prediction results of simple neural networks were better than that of conventional Black-sholes. A hybrid neural network was designed that employed some aspects of Black-Scholes model. The hybrid neural network outperformed the tradition forecasting model and improved prediction results of simple neural networks.