A Reproducible Baseline for Forecasting High-Frequency Realized Volatility with Order-Flow Features
Journal of Economic Theory and Business Management · 2025
DOI: 10.70393/6a6574626d.333538 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of Economic Theory and Business Management, 2025. Verfasst von Zhangqi Liu. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
This paper proposes an interpretable and reproducible baseline model for predicting high-frequency realized volatility based on limit order book numbers, balancing methodological rigor with the requirements of regulated financial practices. We combine economically motivated covariates with a monotonic constrained gradient boosting model that encodes directional prior information based on microstructure theory. The evaluation scheme integrates rolling windows and time-constrained k-fold cross-validation to assess its cross-domain performance. Nevertheless, some sensitivity to mechanism shifts may remain, and excluding news or cross-platform signals may limit coverage, indicating a need for further research.
Zitieren
APA
Liu, Z. (2025). A Reproducible Baseline for Forecasting High-Frequency Realized Volatility with Order-Flow Features. Journal of Economic Theory and Business Management. https://doi.org/10.70393/6a6574626d.333538
Vancouver
Liu Z. A Reproducible Baseline for Forecasting High-Frequency Realized Volatility with Order-Flow Features. Journal of Economic Theory and Business Management. 2025. doi:10.70393/6a6574626d.333538
BibTeX
@article{Liu2025,
author = {Liu, Zhangqi},
title = {A Reproducible Baseline for Forecasting High-Frequency Realized Volatility with Order-Flow Features},
journal = {Journal of Economic Theory and Business Management},
year = {2025},
doi = {10.70393/6a6574626d.333538}
}
EUVIMED