Volatility Forecasting with Reinforcement Learning-Augmented GARCH Models: A Hybrid Framework Using PLM-GARCH and FIGARCH

This study develops a novel hybrid framework for financial volatility forecasting by integrating Reinforcement Learning with GARCH models. Within this framework, the PLM-GARCH and FIGARCH models are employed to capture long-term and periodic volatility patterns, while reinforcement learning algorithms are utilized to potentially enhance predictive performance. The integration of these two approaches results in an adaptive model aimed at effectively simulating the complex behavior of financial time series. Empirical results suggest that the proposed framework may improve forecasting accuracy compared to traditional models in financial volatility and Value-at-Risk (VaR) estimation. This research contributes to the development of GARCH-type models through artificial intelligence integration and highlights possible practical applications in financial risk management and market volatility forecasting, particularly under complex and rapidly changing market conditions. The study covers analysis on datasets including the S&P ۵۰۰, NASDAQ ۱۰۰, and Bitcoin/USD daily returns.

کلیدواژه ها:

Volatility forecasting ، Reinforcement Learning ، PLM-GARCH ، FIGARCH ، Risk Management