非线性时间序列的信息流研究

The Research on Information Flow in Nonlinear Time Series

基于Takens定理的相空间重构为利用时间序列分析研究非线性动力系统提供了理论基础。相空间重构不仅给出了用于分析的动态模型,而且揭示了存在于状态变量之间的信息流。本文以Lorenz模型为例,通过相空间重构,验证了存在于Lorenz模型各状态变量之间的信息流。提出了利用信息流概念进行虚拟测量的方法。

Based on Takens' phase space reconstruction theorem, it is possible for analyzing and studying nonlinear dynamics system using time series. By means of phase space reconstruction, we not only obtain the dynamic mathematics model for original system, but also get knowledge of Information flow in the variables of dynamic system. In this paper, we verify the information flow in the Lorenz model. A new method of virtual measurement using information flow is put forward.