Volume 2,Issue 7
基于自适应权重的MIDAS-GARCH 模型
针对传统MIDAS-GARCH 模型中固定权重函数难以有效捕捉高频数据微观结构噪声与时变特征的局限，本文提出自适应权重优化的MIDAS-GARCH 模型。通过引入基于高频数据交易特征与波动动态调整权重的函数，克服传统方法权重分配模式固定的局限，实现高频信息的高效融合。理论层面论证了模型参数估计量的一致性与渐近正态性，并采用拟极大似然估计（QMLE）实现参数求解。模拟研究和实证研究验证了模型在不同数据生成机制下的高估计精度与稳健性。
[1]Engle R F. Autoregressive Conditional Heteroscedasticity with Estimates of the Variance of U.K. Inflation[J]. Econometrical, 1981, 50(04): 987-1008.
[2]Bollerslev, Tim. Generalized autoregressive conditional heteroskedasticity[J]. Economics and Econometrics Research Institute (EERI), 1986, (31): 307-327
[3]Taylor S J. Modelling financial time series[J]. John Wiley & Sons, 1986: 296.
[4]Zhang, L., Mykland, P. A., & Aït-Sahalia, Y. A Tale of Two Time Scales: Determining Integrated Volatility With Noisy High-Frequency Data[J]. Journal of the American Statistical Association, 2005(100), 1394–1411.
[5]Barndorff ‐Nielsen O. E, Reinhard H P, Lunde A, et al. Realized kernels in practice: trades and quotes[J]. Econometrics Journal, 2009, 12(03): C1–C32.
[6]Asgharian H, Hou A J, Javed F. The Importance of the Macroeconomic Variables in Forecasting Stock Return Variance: A GARCH ‐MIDAS Approach[J]. Journal of Forecasting, 2013, (32): 600-612.
[7] 郑挺国, 尚玉皇. 基于宏观基本面的股市波动度量与预测[J]. 世界经济,2014(12):118-139.
[8]Engle R F, Ghysels E, Sohn B. Stock Market Volatility and Macroeconomic Fundamentals [J]. Review of Economics Statistics,2013,(95):776797.
[9]ALMON S. The Distributed Lag Between Capital Appropriations and Expenditures[J]. Econometrica, 1965, 33: 178-196.
[10]Feng, W., Zhang, X., Chen, Y., & Song, Z. Linear regression estimation using intraday high frequency data[J]. AIMS Mathematics, 2023, 8(06): 13123-13133.
[11]Han H, Park J Y. ARCH/GARCH with persistent covariate: Asymptotic theory of MLE[J]. Journal of Econometrics, 2012, 167(01): 95-112.
[12]Buhlmann P. SieveBootstrap for Time Series[J]. Bernoulli, 1997, (03): 123-148.
[13] 李莉丽，张兴发，李元，等． 基于高频数据的日频 GARCH 模型估计[J]. 广西师范大学学报( 自然科学版) ，2021，39( 4) : 68-78．
[14] 许敏. 基于自适应权重函数的MIDAS-GARCH 模型及其应用研究[D]. 广州：广州大学, 2025. DOI: 10.27040/d.cnki.ggzdu.2025.001976.