时间序列条件矩持续和协同持续性研究  

Study of persistence and copersistence for conditional moments of time series

作　　者：李松臣[1] 张世英[2] 

机构地区：[1]深圳大学数学与计算科学学院,广东深圳518060 [2]天津大学管理学院,天津300072

出　　处：《系统工程学报》2008年第2期129-136,167,共9页Journal of Systems Engineering

基　　金：国家自然科学基金资助项目(70471050)

摘　　要：根据非线性单整理论提出了矩持续和矩协同持续的概念,从时间序列各阶矩长期均衡关系的角度构建了协整和波动协同持续性的统一数学框架,并进一步给出了矩持续和矩协同持续性在三阶矩和四阶矩情形的表现形式;其次在 ARMA-GARCHSK 模型体系下给出了时间序列矩持续和矩协同持续存在的充分必要条件;最后给出了向量 ARMA-GARCHSK 模型矩协同持续的 ARMA 表示形式和误差修正模型.Persistence and copersistence in conditional moments of time series is proposed based on the concept of nonlinear cointegration, an uniform mathematical frame is constructed to study cointegration and copersistence in variance as an integration, and the expressions of persistence and copersistence in conditional skewness and kurtosis are given as extended forms. Then the sufficient and necessary conditions for persistence and copersistence in conditional moments of ARMA-GARCHSK model are studied, and the ARMA expression and error correction expression of this model are given finally.

关 键 词：时间序列 非线性单整 条件矩协同持续 ARMA—GARCHSK模型 误差修正模型 

分 类 号：F224.0[经济管理—国民经济]