Black-Scholes versus Artificial Neural Networks in Pricing Call Warrants: the Case of China Market

The back-propagation neural network is used to pricing call warrants, and the input variables of network model are investigated. The market call warrants prices quoted on Shanghai stock exchange and Shenzhen stock exchange are used to train and simulate the network model. The results show that the performances of the proposed network model produce better call warrant prices than Black-Scholes, and better depict the price characteristics of China's call warrants. The pricing error of Black-Scholes is detailed analyzed, and the market particularities of China's call warrants with different contract terms and price characteristics are also discussed.