Разработка торговой стратегии на основе прогнозирования ценового движения с использованием GARCH-модели и оптимизации через LSTM

Abstract

Predicting the dynamics of financial assets is a crucial task in developing effective trading strategies. In this paper, we apply the GARCH model to estimate market volatility and employ an LSTM neural network to optimize trade execution. The effectiveness of the strategy is evaluated using historical data, with a focus on risk-adjusted returns. The results are assessed through key financial metrics, including the Sharpe ratio. The implementation is performed in Python (arch library), leveraging modern machine learning and statistical modeling libraries