基于改进ICA模型的高维波动率估计

• 作 者: 赵丽丽 张波
• 机 构: 中国人民大学应用统计科学中心统计学院,
• 出 处: 数理统计与管理
• 发表日期: 2017年01期
• 关键词: volatility multidimensional GARCH model ICA CIC-GARCH

本文对IC-GARCH模型进行改进,放宽了ICA关于独立成分是IID的假设,考虑独立成分为ARMA模型的平稳过程,提出了基于自相关结构的IC-GARCH估计方法,并给出了估计量的理论性质和计算效率较高的迭代估计算法。最后,对改进方法进行了模拟和实证分析,结果表明本文提出的模型能够使多元GARCH模型应用于高维金融数据,并大大提高了金融资产收益波动率的估计精度。

This paper improves the IC-GARCH model, relaxes the ICA assumption that the independent component is IID, considers the independent component as the stationary process of the ARMA model, proposes an IC-GARCH estimation method based on the autocorrelation structure, and gives the estimated Iterative estimation algorithm with high theoretical efficiency and computational efficiency. Finally, the simulation and empirical analysis of the improved methods show that the model proposed in this paper can make the Multivariate GARCH model applied to high-dimensional financial data and greatly improve the estimation accuracy of the volatility of the return on financial assets.