基于混频数据和深度学习模型的股市波动率预测

doi:10.6040/j.issn.1671-9352.0.2025.119

摘要/Abstract

摘要： 面对股市波动中影响要素多源、频率多样、关系非线性等潜在挑战,选取上证综指作为研究对象,构建GARCH-MIDAS模型与深度学习模型CNN-BiLSTM-Attention的组合框架,并将月度宏观经济指标、日度经济政策不确定性以及股票交易技术指标等预测因子引入其中,对上证综指波动率进行建模。实证结果表明: 相比于其他基准模型, CNN-BiLSTM-Attention模型在股市波动率预测中表现优异,样本外预测损失函数值相较于其他对比模型均处于最低水平,其中均方误差相较LSTM模型降低47%;此外宏观经济指标等多种预测因子的加入也发挥了关键作用,为股市波动率预测提供有价值的信息,增强模型预测效果。

关键词: 波动率预测, 宏观经济, GARCH-MIDAS模型, CNN-BiLSTM-Attention模型

Abstract: To address challenges in stock market volatility prediction, such as multiple influencing factors, mixed data frequencies, and nonlinear relationships, the Shanghai Composite Index is selected as the modeling target, and a hybrid framework is constructed by combining the GARCH-MIDAS model with the CNN-BiLSTM-Attention deep learning model. Monthly macroeconomic indicators, daily economic policy uncertainty, and technical trading indicators are introduced as predictive variables. Empirical results show that the CNN-BiLSTM-Attention model outperforms benchmark models, achieving the lowest out-of-sample prediction loss. In particular, its mean squared error is reduced by 47 percent compared to the LSTM model. The inclusion of macroeconomic and other predictors also provides valuable information, enhancing forecasting performance.

Key words: volatility forecasting, macroeconomics, GARCH-MIDAS model, CNN-BiLSTM-Attention model

中图分类号:

O211

张玉凤,张同辉,王元月. 基于混频数据和深度学习模型的股市波动率预测[J]. 《山东大学学报(理学版)》, 2026, 61(9): 96-107.

ZHANG Yufeng, ZHANG Tonghui, WANG Yuanyue. Volatility forecasting of stock market based on deep learning model with mixed-frequency data[J]. JOURNAL OF SHANDONG UNIVERSITY(NATURAL SCIENCE), 2026, 61(9): 96-107.

参考文献

[1] Bollerslev T. Generalized autoregressive conditional heteroskedasticity[J]. Journal of Econometrics, 1986, 31(3): 307-327.
[2] Andersen T G, Bollerslev T. Answering the skeptics: yes, standard volatility models do provide accurate forecasts[J]. International Economic Review, 1998, 39(4): 885-905.
[3] Corsi F. A simple approximate long-memory model of realized volatility[J]. Journal of Financial Econometrics, 2009, 7(2): 174-196.
[4] Shu Qi, Xiong Heng, Jiang Wenjun, et al. A novel perspective on forecasting non-ferrous metals volatility: integrating deep learning techniques with econometric models[J]. Finance Research Letters, 2023, 58: 104482.
[5] Zhou Junhao, He Zhanhong, Song Yanan, et al. Precious metal price prediction based on deep regularized self-attention regression[J]. IEEE Access, 2020, 8: 2178-2187.
[6] Kim H Y, Won C H. Forecasting the volatility of stock price index: a hybrid model integrating LSTM with multiple GARCH-type models[J]. Expert Systems with Applications, 2018, 103(8): 25-37.
[7] Song Yuping, Tang Xiaolong, Wang Hemin, et al. Volatility forecasting for stock market incorporating macroeconomic variables based on GARCH-MIDAS and deep learning models[J]. Journal of Forecasting, 2023, 42(1): 51-59.
[8] 汪刘凯, 张小波, 闫相斌, 等. 基于混频数据驱动神经网络模型的波动率预测研究[J]. 系统工程理论与实践, 2023, 43(12): 3488-3504. Wang Liukai, Zhang Xiaobo, Yan Xiangbin, et al. Research on volatility prediction based on mixed frequency data-driven neural network model[J]. Systems Engineering-Theory & Practice, 2023, 43(12): 3488-3504.
[9] 杨科, 张洲深, 田凤平. 高频数据环境下我国股票市场的波动率预测: 基于机器学习和HAR模型的融合研究[J]. 计量经济学报, 2023, 3(3): 886-904. Yang Ke, Zhang Zhoushen, Tian Fengping. The volatility forecasting of Chinese stock market under high-frequency data environment: fusion research based on machine learning and HAR model[J]. China Journal of Econometrics, 2023, 3(3): 886-904.
[10] Mutinda J K, Langat A K. Stock price prediction using combined GARCH-AI models[J]. Scientific African, 2024, 26: e02374.
[11] Nonejad N. Forecasting aggregate stock market volatility using financial and macroeconomic predictors: which models forecast best, when and why? [J]. Journal of Empirical Finance, 2017, 42: 131-154.
[12] Barro R J. Rare disasters and asset markets in the twentieth century[J]. Quarterly Journal of Economics, 2006, 121(3): 823-866.
[13] 刘凤根, 吴军传, 杨希特, 等. 基于混频数据模型的宏观经济对股票市场波动的长期动态影响研究[J]. 中国管理科学, 2020, 28(10): 65-76. Liu Fenggen, Wu Junchuan, Yang Xite, et al. Long-run dynamic effect of macro-economy on stock market volatility based on mixed frequency data model[J]. Chinese Journal of Management Science, 2020, 28(10): 65-76.
[14] Fang Libing, Yu Honghai, Xiao Wen. Forecasting gold futures market volatility using macroeconomic variables in the United States[J]. Economic Modelling, 2018, 72: 249-259.
[15] Yi Adan, Yang Menglong, Li Yongshan. Macroeconomic uncertainty and crude oil futures volatility-evidence from China crude oil futures market[J]. Frontiers in Environmental Science, 2021, 9: 636903.
[16] Huang Y, Paul L. Measuring economic policy uncertainty in China[J]. China Economic Review, 2020, 59: 101367.
[17] Alaali F. The effect of oil and stock price volatility on firm level investment: the case of UK firms[J]. Energy Economics, 2020, 87: 104731.
[18] Li Tao, Ma Feng, Zhang Xuehua, et al. Economic policy uncertainty and the Chinese stock market volatility: novel evidence[J]. Economic Modelling, 2020, 87: 24-33.
[19] Williams G. Is US trade policy uncertainty powerful enough to predict global output volatility?[J]. Journal of International Trade & Economic Development, 2021, 30(1): 138-154.
[20] 龚旭, 林伯强. 跳跃风险、结构突变与原油期货价格波动预测[J]. 中国管理科学, 2018, 26(11): 11-21. Gong Xu, Lin Boqiang. Jump risk, structural breaks and forecasting crude oil futures volatility[J]. Chinese Journal of Management Science, 2018, 26(11): 11-21.
[21] Bates D S. How crashes develop: intradaily volatility and crash evolution[J]. The Journal of Finance, 2019, 74(1): 193-238.
[22] 陈逢文, 金启航, 胡宗斌. 中国股市价格跳跃行为的验证及应用[J]. 财贸经济, 2018, 39(9): 74-88. Chen Fengwen, Jin Qihang, Hu Zongbin. Research on improvement of Fama-French three factor model based on jump behaviour[J]. Finance & Trade Economics, 2018, 39(9): 74-88.
[23] Caporin M. The role of jumps in realized volatility modeling and forecasting[J]. Journal of Financial Econometrics, 2023, 21(4): 1143-1168.
[24] Engle R F, Ghysels E, Sohn B. Stock market volatility and macroeconomic fundamentals[J]. Review of Economics & Statistics, 2013, 95(3): 776-797.
[25] Hansen P R, Lunde A, Nason J M. The model confidence set[J]. Econometrica, 2011, 79(2): 453-497.
[26] Hansen P R. A test for superior predictive ability[J]. Journal of Business & Economic Statistics, 2005, 23(4): 365-380.

相关文章 15

[1]	段颖鹏, 胡琳. 带泊松跳随机微分方程的组合解法[J]. 《山东大学学报(理学版)》, 2026, 61(4): 123-132.
[2]	吕平涛,王才士,赵积君. 基于Hadamard游荡的一类具有相同特征值和连续谱的两态量子游荡[J]. 《山东大学学报(理学版)》, 2024, 59(8): 84-93.
[3]	高玉峰,冯德成. 条件弱鞅和条件N-弱鞅的上(下)穿不等式[J]. 《山东大学学报(理学版)》, 2024, 59(6): 122-126.
[4]	张亚茹,夏莉,张典秋. 混合双分数布朗运动下的永久美式回望期权定价[J]. 《山东大学学报(理学版)》, 2024, 59(4): 98-107.
[5]	严晓东. 策略极限理论与策略统计学习[J]. 《山东大学学报(理学版)》, 2024, 59(1): 1-10, 45.
[6]	徐诚浩,王开永. 二维离散时间风险模型的破产概率[J]. 《山东大学学报(理学版)》, 2023, 58(11): 27-34, 52.
[7]	杨鹏,张晓燕. 基于Stackelberg博弈的最优再保险合同设计[J]. 《山东大学学报(理学版)》, 2023, 58(11): 1-14,26.
[8]	张诗苗,吕艳. 带Lévy噪声的Lotka-Volterra竞争模型的参数估计[J]. 《山东大学学报(理学版)》, 2023, 58(10): 24-31.
[9]	丁佳鑫,郭永峰,米丽娜. 高斯噪声和Lévy噪声激励下欠阻尼周期势系统的相变行为[J]. 《山东大学学报(理学版)》, 2023, 58(8): 111-117.
[10]	高琦,戴洪帅,武艳华. 基于MPEWMA控制图的串联排队网络的监测与控制[J]. 《山东大学学报(理学版)》, 2023, 58(8): 104-110, 117.
[11]	陈叶君,丁惠生. 带有Stepanov概周期系数的无穷维随机微分方程的θ-概周期解[J]. 《山东大学学报(理学版)》, 2023, 58(6): 113-126.
[12]	周玉兰,李转,李晓慧. 连续时间Guichardet-Fock空间中修正随机梯度算子的性质[J]. 《山东大学学报(理学版)》, 2018, 53(12): 62-68.
[13]	李永明,聂彩玲,刘超,郭建华. 负超可加阵列下非参数回归函数估计的相合性[J]. 《山东大学学报(理学版)》, 2018, 53(12): 69-74.
[14]	陈昊君,郑莹,马明,边莉娜,刘华. 自激滤过泊松过程的协方差[J]. 《山东大学学报(理学版)》, 2018, 53(12): 75-79.
[15]	马明,边莉娜,刘华. 基于事件点联合分布的自激滤过泊松过程的低阶矩[J]. 山东大学学报（理学版）, 2018, 53(4): 55-58.

多维度评价

Viewed

Full text

Abstract

Cited

Shared

Discussed