METHOD FOR PREDICTING NO ARBITRAGE OPTION PRICE

摘要

Selecting a parametric model and a neural network model to be trained according to the parameter model as a model representing the stock market; Estimating a parameter that enables the parameter model to reflect a real stock market for option trading by using actual transaction data of options collected in the stock market; When the maximum and minimum values of the option expiration that require price prediction, and the maximum and minimum values of the option price range are determined, generate two-dimensional grid points between the maximum and minimum values of the option expiry and the option price range, and estimate the parameter Generating virtual option data corresponding to each of the two-dimensional grid points according to a method of determining a non-arbitrary option price in the completed parameter model; And learning the selected machine learning model using the virtual option data and the actual transaction data, and predicting a margin-free option price for new input data using the trained machine learning model. A prediction method is provided.