Using Neural Networks to Improve Financial Option Pricing

This study explores using neural networks (NN) as an alternative to traditional Black-Scholes pricing for financial options. By training a backpropagation neural network using Black-Scholes values as inputs, the model better predicts market option prices, especially for in-the-money and out-of-the-money contracts. The results suggest NNs can adjust for real-world market imperfections that Black-Scholes cannot. This hybrid approach offers a promising new direction for option valuation.