Autoencoder Enhanced Realised GARCH on Volatility Forecasting

Paper: arXiv 2411.17136

Authors: Qianli Zhao, Chao Wang, Richard Gerlach, Giuseppe Storti, Lingxiang Zhang

Abstract

Realised volatility has become increasingly prominent in volatility forecasting due to its ability to capture intraday price fluctuations. With a growing variety of realised volatility estimators, each with unique advantages and limitations, selecting an optimal estimator may introduce challenges. In this thesis, aiming to synthesise the impact of various realised volatility measures on volatility forecasting, we propose an extension of the Realised GARCH model that incorporates an autoencoder-generated synthetic realised measure, combining the information from multiple realised measures in a nonlinear manner. Our proposed model extends existing linear methods, such as Principal Component Analysis and Independent Component Analysis, to reduce the dimensionality of realised measures. The empirical evaluation, conducted across four major stock markets from January 2000 to June 2022 and including the period of COVID-19, demonstrates both the feasibility of applying an autoencoder to synthesise volatility measures and the superior effectiveness of the proposed model in one-step-ahead rolling volatility forecasting. The model exhibits enhanced flexibility in parameter estimations across each rolling window, outperforming traditional linear approaches. These findings indicate that nonlinear dimension reduction offers further adaptability and flexibility in improving the synthetic realised measure, with promising implications for future volatility forecasting applications.

Complexity vs Empirical Score

• Math Complexity: 6.5/10
• Empirical Rigor: 8.0/10
• Quadrant: Holy Grail — high math complexity, high empirical rigor

Why this score: The paper introduces an innovative nonlinear dimension-reduction framework using autoencoders to combine multiple realized volatility estimators into a Realised GARCH model. It features a robust empirical study across four major global equity markets over a 22-year period (including the COVID-19 shock), demonstrating clear out-of-sample forecasting gains over standard PCA/ICA approaches.

Research Flowchart

  flowchart TD
    A[Research Goal: Synthesize multiple realized volatility measures to improve Realised GARCH forecasting] --> B[Input Data: Intraday prices from 4 major stock markets 2000–2022]
    B --> C[Dimensionality Reduction: Compare Linear PCA/ICA vs. Nonlinear Autoencoders]
    C --> D[Model Construction: Integrate Autoencoder synthetic measure into Realised GARCH]
    D --> E[Evaluation: One-step-ahead rolling volatility forecasting including COVID-19 period]
    E --> F[Outcomes: Autoencoder model outperforms linear methods with superior flexibility and accuracy]

Cited References (277)

1. The model confidence set — link
2. Quasi-maximum likelihood estimation and inference in dynamic models with time-varying covariances — link
3. Adam: A method for stochastic optimization — link
4. Archived: WHO timeline - COVID -19 — link
5. A long short-term memory enhanced realized conditional heteroskedasticity model — link
6. Improving the accuracy of tail risk forecasting models by combining several realized volatility estimators — link
7. Extended realized GARCH models — link
8. Multiple measures realized GARCH models — link

3 of 277 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2518 linked papers.