A hybrid model for carbon prices: Integrating investor attention, mixed-frequency data, quantile regression and deep learning

Key Points

• The model integrates investor attention and deep learning techniques, enhancing prediction accuracy of carbon prices.
• Key evidence shows a significant correlation between investor attention metrics and carbon pricing fluctuations over recent years.
• Using quantile regression, this analysis allows for detecting varying impacts of investor sentiment across different price levels.
• These findings indicate potential new strategies for predicting market trends, but further validation in real-world scenarios is necessary.