SpotV2Net: Multivariate Intraday Spot Volatility Forecasting via Vol-of-Vol-Informed Graph Attention Networks

ArXiv ID: 2401.06249 “View on arXiv”

Authors: Unknown

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.

Keywords: Graph Attention Network, Intraday Volatility Forecasting, Fourier Estimates, Spillover Effects, GNNExplainer, Equities

Complexity vs Empirical Score

• Math Complexity: 7.5/10
• Empirical Rigor: 8.0/10
• Quadrant: Holy Grail
• Why: 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.

  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"]