基于蒙特卡洛技术时变Copula的比较

Comparison of Time - varying Copula Based on Monte Carlo Technology

从Kendallτ、尾相关系数出发建立了两种时变Copula模型。基于Copula理论,证实了从Kendallτ建立的时变Copula具有明显优于从尾相关建立时变Copula的特征。利用蒙特卡洛技术,验证了Kendallτ建立的时变Copula更适合金融数据的相依关系的描述。

Two variable time-varying Copula models were established based on Kendall τ and tail dependence.It was confirmed that a time-varying Copula established from Kendall τ had obviously better features than the tail dependence based on the theory of Copula.Using Monte Carlo Methods,it was verified that time varying Copula Model based on Kendall τ was much more suitable for dependency mechanism analysis of financial data than its counterpart based on tail dependence.