Copula相依序列与Copula自回归模型探讨  

Discussion on Copula Dependence Series and Copula Auto-Regression Model

作　　者：李述山 Li Shushan(College of Mathematics and Systems Science, Shandong University of Science and Technology, Qingdao Shandong 266590, Chin)

机构地区：[1]山东科技大学数学与系统科学学院

出　　处：《统计与决策》2018年第11期23-27,共5页Statistics & Decision

基　　金：国家自然科学基金资助项目(11271007)

摘　　要：文章提出了Copula相依序列的概念,讨论了Copula相依序列的性质及与严平稳序列之间的关系。建立了一个时间序列的非线性模型——Copula自回归模型,给出了相应的参数估计方法、相依阶数的确定方法。并基于Copula自回归模型给出了一种点预测方法、一种区间预测方法以及一个时变Va R的估计方法。运用算例说明了Copula自回归模型及相关方法的有效性。This paper presents the concept of copula dependence series and discusses the properties of copula dependence series and the relation between copula dependence series and strict stationary series. A new non-linear model about time series named copula auto-regression model（CAR） is established, and the method of parameter estimate and the determination method of the dependence order of copula auto-regression model is given. And based on copula auto-regression model, the paper also puts forward the point prediction method and interval prediction method as well as estimate method of time-varying Va R. Finally a practical example is employed to illustrate the validity of the model and proposed method.

关 键 词：Copula相依序列 Copula自回归模型 点预测 区间预测 时变VaR 

分 类 号：O212[理学—概率论与数理统计]