Stock Market Volatility Forecasting Based on Support Vector Machine and GARCH Model

Jiawei Zhang

doi:10.47297/wspceWSP2516-251919.20250903

PDF
Export
Share
Collection
Album

Stock Market Volatility Forecasting Based on Support Vector Machine and GARCH Model

Volume 9, Issue 3

Views: 0 Downloads: 0

Stock Market Volatility Forecasting Based on Support Vector Machine and GARCH Model
Stock Market Volatility Forecasting Based on Support Vector Machine and GARCH Model
Economics collageUniversity of BristolBristol, BS8 1QU, UK
Vol. 9, Issue 3, Pages: 96-100(2025)
Published：30 June 2025
DOI：10.47297/wspceWSP2516-251919.20250903
Accepted：

Scan QR Code

Zhang Jiawei.Stock Market Volatility Forecasting Based on Support Vector Machine and GARCH Model[J].Creative Economy,2025,09(03):96-100.
DOI：

Zhang Jiawei.Stock Market Volatility Forecasting Based on Support Vector Machine and GARCH Model[J].Creative Economy,2025,09(03):96-100. DOI： 10.47297/wspceWSP2516-251919.20250903.

摘要

Abstract

Stock market volatility serves as a crucial indicator for assessing financial risk and gauging the level of uncertainty present in the market. The traditional GARCH (Generalized Autoregressive Conditional Heteroskedasticity) model has demonstrated considerable efficacy in modeling the phenomenon of volatility clustering observed in time series data. However

this model encounters significant limitations when it comes to addressing the complexities associated with nonlinear and high-dimensional data sets. In contrast

the Support Vector Machine (SVM)

a well-established statistical learning technique

excels in managing scenarios characterized by small sample sizes

high dimensionality

and nonlinear relationships. Recognizing the complementary strengths of these approaches

this paper introduces a novel hybrid forecasting model that seamlessly integrates the SVM methodology with the GARCH framework

aiming to enhance the precision of stock market volatility predictions.To validate the efficacy of this proposed hybrid model

we utilize daily return data from the Shanghai Composite Index and the Shenzhen Component Index

spanning the period from 2017 to 2023. An empirical comparison is conducted between the SVM-GARCH hybrid model and the traditional GARCH model

as well as a standalone SVM model. The empirical results reveal that the hybrid model consistently outperforms the individual models across multiple performance metrics

including predictive accuracy

robustness

and directional correctness. These findings underscore the hybrid model's strong practical applicability and highlight its potential for broader implementation in financial markets.

关键词

Keywords

references

Chen S . and Yu S . Risk analysis of the Shenzhen and Shanghai stock markets using the VaR method based on GARCH models with different distribution assumptions [M ] . Systems Engineering Theory and Practice . 2002 .

Zheng W . The concept of VaR and its application in China [J ] . Journal of International Finance Studies . 1997 .

Wang C. , Wan , H. Li, G. Estimating financial market risk VaR based on MCMC [J ] . Journal of Management Science . 2000 .

Wu , S. Chen, B. Theoretical and empirical study on risk measurement methods and financial asset allocation models [J ] . Economic Research Journal . 1999 .

Yang X . and Xia Y . VaR research on open-end fund portfolios based on the Copula method [J ] . Systems Engineering . 2008 .

Liu B . Research on financial distress prediction of commercial banks based on integrated Random Forest and SVM [J ] . Mathematics in Practice and Theory . 2020 .

Rubio Lihki ; Palacio Pinedo Adriana ; Mejía Castaño Adriana ; Ramos Filipe . Forecasting volatility by using wavelet transform, ARIMA and GARCH models.Journal | [J ] Eurasian Economic Review. Volume 13 , Issue 3- 4 . 2023. PP 803-830.

Banik Shouvik ; Sharma Nonita ; Mangla Monika ; Mohanty Sachi Nandan; S. Shitharth . LSTM based decision support system for swing trading in stock market.Journal | [J ] Knowledge-Based Systems. Volume 239 , Issue . 2022 .

Pedro Correia S . Bezerra; Pedro Henrique M. Albuquerque.Volatility forecasting via SVR–GARCH with mixture of Gaussian kernels.Journal | [J ] Computational Management Science. Volume 14 , Issue 2 . 2017. PP 179-196.

Bildirici M. , & Ersin O . O . Support vector machine GARCH models in modeling conditional volatility: An application to Turkish financial markets [M ] . ICEOS Conference Report . 2013 .

O Chia Chuang ; Rangan Gupta ; Christian Pierdzioch ; Buliao Shu . Financial Uncertainty and Gold Market Volatility: Evidence from a Generalized Autoregressive Conditional Heteroskedasticity Variant of the Mixed-Data Sampling (GARCH-MIDAS) Approach with Variable Selection.Journal | [J ] Econometrics. Volume 12 , Issue 4 . 2024. PP 38-38.

Shiyi Chen ; Wolfgang K . Härdle; Kiho Jeong.Forecasting volatility with support vector machine‐based GARCH model.Journal | [J ] Journal of Forecasting. Volume 29 , Issue 4 . 2010. PP 406-433.

Michańków J. , et al . Combining Deep Learning and GARCH Models for Financial Volatility and Risk Forecasting [M ] . arXiv . 2023 .

Ha Young Kim ; Chang Hyun Won . Forecasting the volatility of stock price index: A hybrid model integrating LSTM with multiple GARCH-type models.Journal | [J ] Expert Systems With Applications. Volume 103 , Issue . 2018 . PP 25- 37 .

Yamaka W. , Srichaikul W. , & Maneejuk P . Support Vector Machine-Based GARCH-type Models: Evidence from ASEAN-5 Stock Markets [M ] . Data Science for Financial Econometrics . 2021 .

Alert me when the article has been cited

Submit

Analysis of the impact effect of population and labour mobility on the high-quality development of regional economy from the perspective of big data -- an empirical analysis based on machine learning support vector machine and stata spatial measurement

Related Author

Naiping Zhou School of Economics, Henan University, Kaifeng Henan China

Related Institution

School of Economics, Henan University, Kaifeng Henan China

⁰