论文部分内容阅读
通过引入厚尾的、自由度参数范围更广的广义误差分布(GED)扩展标准马尔科夫转换多分形模型(MSM),探讨MSM-GED模型的参数估计和波动率预测问题,并利用上证综指日收益数据进行实证分析。实证结果表明,上证综指确实存在多分形性,MSM模型的波动预测能力强于(FI)GARCH模型,尤其是中长期波动率预测,MSM-GED能够提供更准确的波动率预测值。这为资产定价和风险管理提供一种新的波动建模选择。
The standard Markovian transformation multifractal model (MSM) is extended by introducing a thick tail with a wider generalized error distribution (GED) with a wider range of degrees of freedom parameters to explore the parameter estimation and volatility prediction of the MSM-GED model. Refers to the daily earnings data for empirical analysis. The empirical results show that the Shanghai Composite Index does exist the multifractality, and the volatility forecasting ability of the MSM model is stronger than the (FI) GARCH model, especially in the medium and long term volatility forecast, MSM-GED can provide more accurate forecast of the volatility. This provides a new wave of modeling options for asset pricing and risk management.