|
|
Finance 2024
基于数据驱动的DD-EWMA模型在港股市场上的金融风险预测
|
Abstract:
金融风险预测是关于波动性、风险价值(VaR)、期望损失(ES)的预测。本文利用传统的GARCH模型、基于数据驱动的指数加权移动平均(DD-EWMA)模型和数据驱动的神经网络波动率模型来对港股市场进行金融风险预测比较。除此之外，本文还运用了滚动DD-EWMA模糊波动率模型、滚动神经网络模糊波动率模型以及滚动GARCH模型来预测VaR和ES。最后，对滚动预测出的VaR和ES分别进行回溯测试，比较这三个模型优劣。模型证明，与GARCH模型相比，DD-EWMA模型具有无偏性。与利用传统估计量样本标准差来估计波动率相比，DD-EWMA模型具有稳定性。实证也表明，这三种模型预测具有高峰度特性的数据的VaR和ES，DD-EWMA波动率模型预测效果最好。
Financial risk prediction is a prediction of volatility, value at risk (VaR) and expected loss (ES). In this paper, the traditional GARCH model, the data-driven index weighted moving average (DD-EWMA) model and the data-driven neural network volatility model are used to forecast and compare the financial risks of the Hong Kong stock market. In addition, this paper also uses rolling DD-EWMA fuzzy volatility model, rolling neural network fuzzy volatility model and rolling GARCH model to predict VaR and ES. Finally, the VaR and ES predicted by rolling are backtested to compare the advantages and disadvantages of these three models. The model proves that the DD-EWMA model is unbiased compared with the GARCH model. The DD-EWMA model is stable compared with the traditional estimator sample standard deviation. The empirical results also show that the volatility models of VaR, ES and DD-EWMA have the best prediction effect on the data with peak degree characteristics.
| [1] | Ruppert, D. and Matteson, D.S. (2015) Risk Management. In: Springer Texts in Statistics, Springer New York, 553-579. https://doi.org/10.1007/978-1-4939-2614-5_19 |
| [2] | Danielsson, J. (2011) Financial Risk Forecasting: The Theory and Practice of Forecasting Market Risk with Implementation in R and Matlab. Wiley, 76-91. |
| [3] | Thavaneswaran, A., Paseka, A. and Frank, J. (2019) Generalized Value at Risk Forecasting. Communications in Statistics—Theory and Methods, 49, 4988-4995. https://doi.org/10.1080/03610926.2019.1610443 |
| [4] | Liang, Y., Thavaneswaran, A., Zhu, Z., Thulasiram, R.K. and Hoque, M.E. (2020). Data-Driven Adaptive Regularized Risk Forecasting. 2020 IEEE 44th Annual Computers, Software, and Applications Conference (COMPSAC), Madrid, 13-17 July 2020, 1296-1301. https://doi.org/10.1109/compsac48688.2020.00-77 |
| [5] | Thavaneswaran, A., Thulasiram, R.K., Zhu, Z., Hoque, M.E. and Ravishanker, N. (2019). Fuzzy Value-at-Risk Forecasts Using a Novel Data-Driven Neuro Volatility Predictive Model. 2019 IEEE 43rd Annual Computer Software and Applications Conference (COMPSAC), Milwaukee, 15-19 July 2019, 221-226. https://doi.org/10.1109/compsac.2019.10210 |
| [6] | Thavaneswaran, A., Thulasiram, R.K., Frank, J., Zhu, Z. and Singh, M. (2019). Fuzzy Option Pricing Using a Novel Data-Driven Feed Forward Neural Network Volatility Model. 2019 IEEE International Conference on Fuzzy Systems (FUZZ-IEEE), Orleans, 23-26 June 2019, 1-6. https://doi.org/10.1109/fuzz-ieee.2019.8858826 |
| [7] | Lucas, A. and Zhang, X. (2016) Score-Driven Exponentially Weighted Moving Averages and Value-at-Risk Forecasting. International Journal of Forecasting, 32, 293-302. https://doi.org/10.1016/j.ijforecast.2015.09.003 |
| [8] | Hyndman, R., Athanasopoulos, G., Bergmeir, C., et al. (2018) Package Forecast—The Comprehensive R Archive Network. https://cran.r-project.org/web/packages/forecast/forecast.pdf |
| [9] | Lyu, Y., Kong, M., Ke, R. and Wei, Y. (2021) Does Mixed Frequency Information Help to Forecast the Value at Risk of the Crude Oil Market? SSRN Electronic Journal, 29, 139-141. https://doi.org/10.2139/ssrn.3774891 |
