New Realised GARCH model enhances volatility forecasting with nonlinear dimension reduction

A new extension of the Realised GARCH model has been proposed to improve volatility forecasting by synthesizing information from multiple realized volatility measures. This model utilizes an autoencoder for nonlinear dimension reduction, outperforming traditional linear methods like Principal Component Analysis and Independent Component Analysis. Empirical evaluations across four major stock markets, including the COVID-19 period, demonstrate the model's superior effectiveness in one-step-ahead rolling volatility forecasting and enhanced flexibility in parameter estimations. AI

IMPACT This research could lead to more accurate financial risk forecasting by leveraging advanced machine learning techniques for data synthesis and dimension reduction.

RANK_REASON The item is an academic paper submitted to arXiv detailing a new model and methodology. [lever_c_demoted from research: ic=1 ai=0.7]

Read on arXiv cs.LG →

• alphaXiv
• arXiv
• CatalyzeX
• COVID-19
• DagsHub
• Gotit.pub
• Hugging Face
• independent component analysis
• principal component analysis
• Qianli Zhao
• Realised GARCH
• ScienceCast

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →