Innovative Study on Volatility Prediction Model for New Energy Stock Indices

Abstract

Stock market volatility is a pivotal research area in finance, and accurately forecasting stock market volatility has long been a challenge for both academia and practice. The emergence of the new energy industry has drawn widespread attention to new energy stock indices; however, research on forecasting their volatility remains limited. To address this, this paper proposes a deep learning ensemble model based on decomposition optimization for predicting the volatility of new energy stock indices. The model comprises three components: variational mode decomposition (VMD), sparrow search algorithm (SSA), and echo state network (ESN). Initially, this paper employs VMD to decompose the original volatility series of new energy stock indices into multiple subsequences. Subsequently, SSA is utilized to optimize ESN. Finally, the constructed VMD-SSA-ESN model is employed to forecast the volatility of new energy stock indices. Through comparative analysis with other forecasting models, this paper finds that the VMD-SSA-ESN model exhibits significantly better forecasting performance across all selected new energy stock index volatility predictions. The research results indicate that the model constructed in this paper can adequately capture the characteristics of the volatility series, and both VMD and SSA effectively enhance the model’s forecasting accuracy and stability. This study can provide robust support for investment decision-making and risk management in the new energy stock market.

Keywords

• Volatility prediction
• deep learning
• variational mode decomposition
• optimization algorithm
• new energy