METHOD FOR PREDICTING NO ARBITRAGE OPTION PRICE

摘要

The present invention provides a method for predicting a no-arbitrage option price which can predict a high-accuracy option price. The method for predicting a no-arbitrage option price comprises: a step of selecting a parametric model to be used in no-arbitrage option price prediction and a neural network model to be learned in accordance with a corresponding parametric model as models for representing a stock market; a step of using real transaction data of options collected in the stock market to estimate a parameter to allow the parametric model to incorporate a realistic stock market for an option transaction; a step of generating two-dimensional grid points between maximum values and minimum values of option expiration and an option price range if a maximum value and a minimum value of the option expiration requiring price prediction and a maximum value and a minimum value of the option price range are determined, and generating virtual option data corresponding to the two-dimensional grid points in accordance with a no-arbitrage option price determination method in the parameter-estimated parametric model; and a step of using the virtual option data and the real transaction data to train the selected neural network model, and using the trained neural network model to predict a no-arbitrage option price for new input data.