Artificial intelligence system for predicting the price of the SIU8 basic asset of the SI-9.18 option in the “long straddle” strategy

Abstract

The features of the “long straddle” option strategy for the Si-9.18 option that was chosen as an object of the research on the Moscow MOEX stock exchange have been investigated. The underlying asset of the Si-9.18 option is a futures SiU8 contract for the US dollar on a 15-min timeframe. The theoretical foundations of option pricing using the Cox-Ross-Rubinstein model have been considered. Based on the Russian Deductor platform, the Kohonen map neural network has been developed to obtain a forecast for the price of the option base asset. There has been suggested and proven a hypothesis that the neural network makes it possible to obtain a price forecast for the SiU8 base asset of a synthetic option in the “long straddle” strategy. Since this option strategy is aimed at making a profit from an intensive movement and increase in volatility, the use of the developed artificial intelligence system based on the “long straddle” strategy will reduce risks and ensure its hedging. The relevance of the study lies in the fact that in the conditions of increase in volatility and risks in conducting trading operations, the application of the predicted values of the neural network that uses the changing parameters of the “option board” is of great practical importance. The study used options (PUT and CALL) taken from the Si-9.18 “option board” of the QUIK trading platform. According to the calculated data of the neural network, the STRIKE_OUT rate was 0.485875706214689 with a STRIKE_ERR of 2.3096796705%.