Integrating GJR-GARCH and LightGBM Quantile Regression Model Based on Gating Mechanism

DOI

10.2991/978-94-6239-801-6_156How to use a DOI?

Keywords

GJR-GARCH Model; LightGBM Quantile Regression; Gating Mechanism Fusion; Cornish Fisher Expansion; Conditional Probability Distribution

Abstract

In the financial field, accurate prediction of price changes and risks is an important reference for making decisions and now the combination of machine learning and statistical models has shown advantages. This study combines dynamic regulation based on market environmental factors with the Glosten-Jagannathan-Runkle Generalized Autoregressive Conditional Heteroskedasticity model (GJR-GARCH) modified by Cornish Fisher expansion and the Light Gradient Boosting Machine (LightGBM) quantile regression machine learning model to capture extreme market risks and construct conditional probability distribution. The experimental results indicate that the hybrid model proposed in this study can effectively capture market risks and improve prediction accuracy, especially in enhancing the prediction accuracy and calibration quality of tail risk (VaR). Solving the problem of traditional models usually using linear autoregressive structures and machine learning facing sparse sample prediction failures, while avoiding the black box effect of deep learning. The result shows that the CRPS value is the lowest in JPM with 0.2148, compared with other models, and it has shown robust performance on representative stocks in different industries with CRPS value between 0.17 to 0.27, proving that the model has good generalization ability.

Copyright

© 2026 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - Lingyan Cai
PY  - 2026
DA  - 2026/10/23
TI  - Integrating GJR-GARCH and LightGBM Quantile Regression Model Based on Gating Mechanism
BT  - Proceedings of the 2026 2nd International Conference on Financial Innovation and Marketing Management (FIMM 2026)
PB  - Atlantis Press
SP  - 1481
EP  - 1491
SN  - 2352-5428
UR  - https://doi.org/10.2991/978-94-6239-801-6_156
DO  - 10.2991/978-94-6239-801-6_156
ID  - Cai2026
ER  -