Paper: arXiv 2401.06249

Abstract

This paper introduces SpotV2Net, a multivariate intraday spot volatility forecasting model based on a Graph Attention Network architecture. SpotV2Net represents assets as nodes within a graph and includes non-parametric high-frequency Fourier estimates of the spot volatility and co-volatility as node features. Further, it incorporates Fourier estimates of the spot volatility of volatility and co-volatility of volatility as features for node edges, to capture spillover effects. We test the forecasting accuracy of SpotV2Net in an extensive empirical exercise, conducted with the components of the Dow Jones Industrial Average index. The results we obtain suggest that SpotV2Net yields statistically significant gains in forecasting accuracy, for both single-step and multi-step forecasts, compared to a panel heterogenous auto-regressive model and alternative machine-learning models. To interpret the forecasts produced by SpotV2Net, we employ GNNExplainer \citep{“ying2019gnnexplainer”}, a model-agnostic interpretability tool, and thereby uncover subgraphs that are critical to a node’s predictions.

Complexity vs Empirical Score

  • Math Complexity: 7.5/10
  • Empirical Rigor: 8.0/10
  • Quadrant: Holy Grail — high math complexity, high empirical rigor

Why this score: The paper involves advanced mathematics including non-parametric Fourier estimators, graph attention network derivations, and multivariate volatility of volatility modeling, while also presenting extensive empirical validation with high-frequency DJIA data and statistical significance testing.

Research Flowchart

  flowchart TD
  A["Research Goal:<br>Multivariate Intraday Spot<br>Volatility Forecasting"] --> B["Data & Inputs<br>Components of Dow Jones<br>Industrial Average"]
  B --> C["Preprocessing<br>Non-parametric High-Frequency<br>Fourier Estimates of Volatility"]
  C --> D["Core Methodology<br>Graph Attention Network<br>with Vol-of-Vol Features"]
  D --> E["Computational Process<br>Spillover Effects &<br>Node/Edge Feature Integration"]
  E --> F["Model Interpretation<br>GNNExplainer<br>Identifying Critical Subgraphs"]
  F --> G["Key Outcomes<br>Significant Gains vs.<br>Panel & ML Benchmarks"]

Cited References (164)

  1. Graph-based methods for forecasting realized covariances (2022)
  2. Volatility forecasting with machine learning and intraday commonality (2023)
  3. Fourier-Malliavin Volatility Estimation: theory and practice (2011)
  4. Graph neural networks for forecasting realized volatility with nonlinear spillover effects (2023)
  5. Multivariate realized volatility forecasting with graph neural network (2022)
  6. Cryptocurrency volatility forecasting using commonality in intraday volatility (2023)
  7. A Similarity-based Approach to covariance forecasting (2023)
  8. Realized volatility forecasting with neural networks (2020)
and 156 more
  1. Novel volatility forecasting using deep learning-long short term memory recurrent neural networks (2019)
  2. Convolutional networks for images, speech, and time series (1995)
  3. Learning internal representations by error propagation (1985)
  4. Long short-term memory (1997)
  5. The graph neural network model (2008)
  6. Semi-supervised classification with graph convolutional networks (2016) — link
  7. How powerful are graph neural networks? (2018) — link
  8. Graph attention networks (2017) — link
  9. Attention is all you need (2017)
  10. Good volatility, bad volatility: Signed jumps and the persistence of volatility (2015)
  11. Forecasting realised volatility using ARFIMA and HAR models (2019)
  12. Cart (1984)
  13. Xgboost: A scalable tree boosting system (2016)
  14. Greedy function approximation: a gradient boosting machine (2001)
  15. Long memory and periodicity in intraday volatility (2015)
  16. Graph neural network for traffic forecasting: A survey (2022)
  17. Spatial-temporal fusion graph neural networks for traffic flow forecasting (2021)
  18. Dynamic spatial-temporal graph convolutional neural networks for traffic forecasting (2019)
  19. Mixgcf: An improved training method for graph neural network-based recommender systems (2021)
  20. Multi-behavior graph neural networks for recommender system (2022)
  21. Graph convolutional neural networks for web-scale recommender systems (2018)
  22. Graph neural networks for social recommendation (2019)
  23. STGSN—A spatial-temporal graph neural network framework for time-evolving social networks (2021)
  24. Convolutional neural networks on graphs with fast localized spectral filtering (2016)
  25. Inductive representation learning on large graphs (2017)
  26. Volatility managed indexes: The importance of intraday data (2021)
  27. Connecting the dots: Multivariate time series forecasting with graph neural networks (2020)
  28. Research on graph neural network in stock market (2022)
  29. A review on graph neural network methods in financial applications (2021) — link
  30. Predicting stock market volatility from candlestick charts: A multiple attention mechanism graph neural network Approach (2022)
  31. An introduction to long-memory time series models and fractional differencing (1980)
  32. A long memory property of stock market returns and a new model (1993)
  33. Combining long memory and level shifts in modeling and forecasting the volatility of asset returns (2018)
  34. From zero-intelligence to queue-reactive: limit-order-book modeling for high-frequency volatility estimation and optimal execution (2023)
  35. Deep learning (2016)
  36. Rectifier nonlinearities improve neural network acoustic models (2013)
  37. The estimation and application of long memory time series models (1983)
  38. Gnnexplainer: Generating explanations for graph neural networks (2019)
  39. Fourier spot volatility estimator: Asymptotic normality and efficiency with liquid and illiquid high-frequency data (2015)
  40. Discovering supply chain links with augmented intelligence (2021) — link
  41. Graph neural networks for modeling causality in international trade (2021)
  42. Bilateral trade modelling with graph neural networks (2020)
  43. Preventing attacks in interbank credit rating with selective-aware graph neural network (2023)
  44. Combining Intra-Risk and Contagion Risk for Enterprise Bankruptcy Prediction Using Graph Neural Networks (2022) — link
  45. An edge feature aware heterogeneous graph neural network model to support tax evasion detection (2023)
  46. Perceiver: General perception with iterative attention (2021)
  47. Efficient deep learning: A survey on making deep learning models smaller, faster, and better (2023)
  48. A comprehensive survey on graph neural networks (2020)
  49. Graph neural networks: A review of methods and applications (2020)
  50. Intraday portfolio risk management using VaR and CVaR: A CGARCH-EVT-Copula approach (2019)
  51. Intraday volatility smile: Effects of fragmentation and high-frequency trading on price efficiency (2021)
  52. Forecasting and trading high-frequency volatility on large indices (2018)
  53. Trading frequency and volatility clustering (2012)
  54. High-Frequency trading strategies (2023)
  55. Forecasting value at risk with intra-day return curves (2020)
  56. Intraday bidirectional volatility spillover across international stock markets: does the global financial crisis matter? (2015)
  57. The intraday volatility spillover index approach and an application in the Brexit vote (2018)
  58. Intra-daily volatility spillovers in international stock markets (2015)
  59. Intraday return dynamics and volatility spillovers between NSE S&P CNX Nifty stock index and stock index futures (2011)
  60. Intradaily periodicity and volatility spillovers between international stock index futures markets (2005)
  61. Intraday price discovery and volatility spillovers in an emerging market (2019)
  62. COVID-induced sentiment and the intraday volatility spillovers between energy and other ETFs (2023)
  63. The influence of intraday seasonality on volatility transmission pattern (2019)
  64. High-frequency volatility co-movements in cryptocurrency markets (2019)
  65. Volatility of volatility estimation: Central limit theorems for the Fourier transform estimator and empirical study of the daily time series stylized facts (2022)
  66. Asymptotic normality for the Fourier spot volatility estimator in the presence of microstructure noise (2022) — link
  67. High-frequency volatility of volatility estimation free from spot volatility estimates (2015)
  68. Estimating covariance via Fourier method in the presence of asynchronous trading and microstructure noise (2011)
  69. Decoupled weight decay regularization (2017) — link
  70. Firm volatility in granular networks (2020)
  71. The common factor in idiosyncratic volatility: Quantitative asset pricing implications (2016)
  72. Multivariate volatility forecasts for stock market indices (2021)
  73. Combining multivariate volatility forecasts: an economic-based approach (2017)
  74. High-dimensional multivariate realized volatility estimation (2019)
  75. The incremental volatility information in one million foreign exchange quotations (1997)
  76. Deep learning for finance: deep portfolios (2017)
  77. Deep learning for portfolio optimization (2020)
  78. Portfolio optimization with return prediction using deep learning and machine learning (2021)
  79. Empirical asset pricing via machine learning (2020)
  80. Deep learning in asset pricing (2023)
  81. Equity2vec: End-to-end deep learning framework for cross-sectional asset pricing (2021)
  82. Deep learning stock volatility with Google domestic trends (2015) — link
  83. Forecasting realized volatility with machine learning: Panel data perspective (2023)
  84. Portfolio selection under non-gaussianity and systemic risk: A machine learning based forecasting approach (2023)
  85. HARNet: A convolutional neural network for realized volatility forecasting (2022) — link
  86. Hats: A hierarchical graph attention network for stock movement prediction (2019) — link
  87. Modeling the momentum spillover effect for stock prediction via attribute-driven graph attention networks (2021)
  88. Time-aware graph relational attention network for stock recommendation (2020)
  89. Financial time series forecasting with multi-modality graph neural network (2022)
  90. Graph correlated attention recurrent neural network for multivariate time series forecasting (2022)
  91. Tabular data: Deep learning is not all you need (2022)
  92. Applying LSTM to time series predictable through time-window approaches (2001)
  93. The performance of LSTM and BiLSTM in forecasting time series (2019)
  94. Fourier series method for measurement of multivariate volatilities (2002)
  95. A Fourier transform method for nonparametric estimation of multivariate volatility (2009)
  96. Asymptotic Normality of the Fourier spot covariance estimator with asynchronous high-frequency prices and an application to spot beta estimation (2024)
  97. The price-leverage covariation as a measure of the response of the leverage effect to price and volatility changes (2022)
  98. Quantitative risk management: Concepts, techniques and tools (2015)
  99. Statistical properties of microstructure noise (2017)
  100. Volatility of volatility and leverage effect from options — → 2305.04137
  101. Volatility of volatility: Estimation and tests based on noisy high-frequency data with jumps (2022)
  102. Forecasting volatility with time-varying leverage and volatility of volatility effects (2020)
  103. A comparison of machine learning methods for predicting the direction of the US stock market on the basis of volatility indices (2023)
  104. The dynamic interrelations of oil-equity implied volatility indexes under low and high volatility-of-volatility risk (2022)
  105. A simple joint model for returns, volatility and volatility of volatility (2023)
  106. Volatility-of-volatility risk in asset pricing (2021)
  107. Volatility-of-volatility risk (2019)
  108. Volatility and correlation forecasting (2006)
  109. Forecasting volatility in financial markets: A review (2003)
  110. An evaluation of volatility forecasting techniques (1996)
  111. How relevant is volatility forecasting for financial risk management? (2000)
  112. Volatility forecasting: The role of lunch-break returns, overnight returns, trading volume and leverage effects (2015)
  113. Realized volatility forecasting and option pricing (2008)
  114. Selecting volatility forecasting models for portfolio allocation purposes (2015)
  115. Forecasting intraday volatility in the us equity market. Multiplicative component GARCH. (2012)
  116. Stock market volatility forecasting: Do we need high-frequency data? (2021)
  117. Forecasting volatility in the financial markets (2011)
  118. The distribution of realized stock return volatility (2001)
  119. Modeling and forecasting realized volatility (2003)
  120. Generalized autoregressive conditional heteroskedasticity (1986)
  121. Autoregressive conditional heteroscedasticity with estimates of the variance of United Kingdom inflation (1982)
  122. A simple approximate long-memory model of realized volatility (2009)
  123. Multivariate GARCH models: A survey (2006)
  124. A capital asset pricing model with time-varying covariances (1988)
  125. Multivariate simultaneous generalized ARCH (1995)
  126. Structured sequence modeling with graph convolutional recurrent networks (2018)
  127. GC-LSTM: Graph convolution embedded LSTM for dynamic network link prediction (2022)
  128. Predicting path failure in time-evolving graphs (2019)
  129. Predictive temporal embedding of dynamic graphs (2019)
  130. Evolvegcn: Evolving graph convolutional networks for dynamic graphs (2020)
  131. T-gcn: A temporal graph convolutional network for traffic prediction (2019)
  132. A3t-gcn: Attention temporal graph convolutional network for traffic forecasting (2021)
  133. Transfer graph neural networks for pandemic forecasting (2021)
  134. Diffusion convolutional recurrent neural network: Data-driven traffic forecasting (2017) — link
  135. Adaptive graph convolutional recurrent network for traffic forecasting (2020)
  136. Spatio-temporal graph convolutional networks: A deep learning framework for traffic forecasting (2017) — link
  137. Chickenpox cases in Hungary: A benchmark dataset for spatiotemporal signal processing with graph neural networks (2021) — link
  138. Algorithms for hyper-parameter optimization (2011)
  139. Random search for hyper-parameter optimization (2012)
  140. Maximum likelihood from incomplete data via the EM algorithm (1977)
  141. The model confidence set (2011)
  142. Comparing predictive accuracy (2002)
  143. Testing the equality of prediction mean squared errors (1997)
  144. Modeling and forecasting large realized covariance matrices and portfolio choice (2017)
  145. Asymptotic theory for a vector ARMA-GARCH model (2003)
  146. The Wishart autoregressive process of multivariate stochastic volatility (2009)
  147. Co-integration and error correction: Representation, estimation, and testing (1987)
  148. Computation of volatility in stochastic volatility models with high-frequency data (2010)
  149. The F ourier- M alliavin volatility ( FMV ol) MATLAB library (2024)
  150. Expected stock returns and variance risk premia (2009)
  151. The volatility of realized volatility (2008)
  152. A machine learning approach to volatility forecasting (2023)
  153. Gradient boosting machines, a tutorial (2013)
  154. The Twitter myth revisited: Intraday investor sentiment, Twitter activity and individual-level stock return volatility (2018)
  155. Early Warning Systems for identifying financial instability (2023)
  156. Volatility forecast comparison using imperfect volatility proxies (2011)

1 of 164 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2518 linked papers.