Stock Index Option Pricing Based On The BP Neural Network Model And The GARCH Model

Option has the functions of price discovery and hedging and is important in the financial market.However,the Black-Scholes pricing model depends on a variety of assumptions.Once the assumptions are not valid in practice,the validity of the model will be affected.Considering other restrictive factors,scholars have made corresponding pricing models,but their pricing effects remain inefficient.In this paper,we study the pricing of CSI 300 index options.To overcome the problems in the parametric models,we use the neural network model to train and learn the option data.Considering the change of volatility,this paper combines the GARCH model with the neural network model and proposes an option pricing method based on the BP neural network model and the GARCH model.Firstly,this paper analyzes the logarithmic returns' distribution of the CSI 300 Index,and obtains the GARCH-t(1,1)model to estimate volatility.Secondly,combining with the estimated volatility,this paper compares the pricing effects of the Black-Scholes pricing model and the BP neural network model.The results show that BP neural network model has a better pricing effect,and when considering the change of underlying asset price volatility,the performance of each model is improved.Thirdly,to overcome the defects of the traditional BP neural network model,this paper employs the particle-swarm-optimization algorithm to improve the model,which further improves the effect of option pricing.Finally,this paper attempts to introduce investor sentiment into the neural network model.The results show that the pricing effect of different neural network models is improved after the introduction of this variable.That is,investor sentiment plays a significant role in option pricing.