一种选取相空间重构最优延迟时间的算法

A New Method to Choose Optimal Delay Time for Phase Space Reconstruction

利用时间序列的可预测性随预测步长变化的规律，提出了一种选取最优延迟时间的新方法；用已知的混沌系统Lorenz、Rossler等产生的时间序列进行仿真计算，与重建扩展法相比，取得了较好的结果，证明了该方法的有效性。

A new method to determine the optimal delay time for the maximum independence of coordinate of phase space is proposed on the basis of predictability of time series. The numerical results show its effectiveness to compute the correlation dimension through phase space reconstruction.