DEFINITIONS AND REPRESENTATIONS OF MULTIVARIATE LONG-RANGE DEPENDENT TIME SERIES
DEFINITIONS AND REPRESENTATIONS OF MULTIVARIATE LONG-RANGE DEPENDENT TIME SERIES
复制标题
DOI：
10.1111/jtsa.12086
复制
发表时间：
2015-01-01
影响因子：
0.9
通讯作者：
Pipiras, Vladas
中科院分区：
文献类型：
--
作者：
Kechagias, Stefanos;Pipiras, Vladas
这里从时域和谱域的角度重新审视多元长程依赖性的概念。自始至终都阐明并强调了所谓的相位参数的作用。特别是，构造了具有一般相位参数的多元长程相关时间序列的因果（单方面）表示的示例。引入了自回归分数积分移动平均序列的多元扩展，并给出了自协方差函数的显式公式。
The notion of multivariate long-range dependence is reexamined here from the perspectives of time and spectral domains. The role of the so-called phase parameters is clarified and stressed throughout. In particular, examples of causal (one-sided) representations of multivariate long-range dependent time series with general-phase parameters are constructed. A multivariate extension of the autoregressive fractionally integrated moving-average series is introduced with explicit formulas for its autocovariance function.