基于蒙特卡洛技术时变Copula的比较  

Comparison of Time - varying Copula Based on Monte Carlo Technology

作　　者：刘娟[1] 王沁[1] 刘军[2] 昌春艳[1] 

机构地区：[1]西南交通大学数学学院,四川成都610031 [2]西南交通大学经济管理学院,四川成都610031

出　　处：《武汉理工大学学报（信息与管理工程版）》2013年第3期336-339,344,共5页Journal of Wuhan University of Technology：Information & Management Engineering

基　　金：教育部人文社会科学研究基金资助项目(09YJCZH104);中央高校基本科研业务费专项资金资助项目(SWJTU12ZT014;SWJ-TU12CX058)

摘　　要：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.

关 键 词：时变COPULA Kendall τ 尾相关 

分 类 号：O212.3[理学—概率论与数理统计]