ПРОГНОЗИРОВАНИЕ ВАРИАЦИОННОГО РЯДА ФЬЮЧЕРСНОГО КОНТРАКТА SIM1 СИСТЕМОЙ ИСКУССТВЕННОГО ИНТЕЛЛЕКТА С ИСПОЛЬЗОВАНИЕМ ДАННЫХ VAR-МОДЕЛИ

The theoretical foundations for assessing the risk / profitability of the variational series of a financial instrument based on the VaR-model are considered. The paper examines some mathematical models for assessing financial risk at a given level of probability, which makes it possible to provide support for making managerial decisions regarding financial risk management when conducting speculative exchange transactions. A hypothesis was put forward and proved that using the developed VaR-model it is possible to estimate the minimum price levels of a financial asset on the next monthly timeframe at a given risk level, and, using the obtained data, calculate the value of the forecast closing price of a futures contract SiM1 based on the generated neural network AI-model. The novelty of the study is that an approach based on the VaR-model is proposed, which allows calculating the predicted values ​​of the closing price of a futures contract SiM1 using a neural network, while in the training sample of the model, in addition to the time series parameters: Po, Ph, Pl, Pc, Volum , also added: profitability, standard deviation, volatility for trading on the derivatives market. The forecasted value was RUB 77,442. and at the current price as of April 30, 2021, the futures contract SiM1 75,665 rubles. you can get 1777 rubles in a month. variation margin per 1 contract. Since the GO (guarantee collateral) for SiM1 is 6156.94 rubles, then the profitability of a futures transaction will be 28.8% (1777 / 6156.94 * 100%).